Supporting Real-Time Applications in an Integrated Services Packet Network: Architecture and Mechanism David D. Clarkl Scott Shenker Lixia Zhang Laboratory for Computer Science Palo Alto Research Center Massachusetts Institute of Technology Xerox Corporation ddc@lcs. mit.edu shenker, lixia@parc.xerox.com Abstract One of the most vexing technical problems that blocks the path towards an ISPN is that of supporting realtim This paper considers the support of real-time applications applications in a packet network. Real-time applications are quite different from standard data applications, and re- review the characteristics of real-time applications. We ob quire service that cannot be delivered wit hin the typical dat a serve that, contrary to the popular view that real-time ap service architecture. In Section 2 we discuss the nature of re a fixed delay bound, some reak here, however, it suffices time applications are dapt to current to observe that one salient characteristic of the realtime network co bound on ture tha While this bound may guaranteed se orm of realbe stat ist ical, in the sense that some small fraction of the time serv he bound itself pre-comp vice archi- which use puting delay bo mechanism that c as well as accomn traffic, an en- two other aspects of a four vice interface and t e by the 1《Introduction《 The current generation of tele phone net w our paper. rent generation of computer networks were each designed to hile the idea of predicted service has been considered be- carry specific and very different kinds of traffic: o our knowledge in ISDN and the increasing applications, this dist ing these sorts of service nclude both fer to here as an Integra the network will would yield a single tele o know when their ing multit ude of adv n of the source's scale, ubiquity of ac ing. There is a broad entify the networking community particular an achievable goal. Ho vision can become a reality. 1Research at MIT was supported by DARPA through NASA Grant g ser- NAG 2-582, by NSF grant NCR-8814187, and by DARPA and NSE through Cooperative Agreement NCR-8919038 with the Corporation the schedul- for National Research Initiatives ing informa eaders. This the rest of the gs to the term. An the fu d is the maximal value of a set of obe cal bounds allow for a certain percentage of violations of the bound; absolute bounds allov
✂✁☎✄✆✄✞✝✠✟☛✡☛☞✍✌✏✎✒✑✔✓✖✕✠✗✍✘✚✙✛☞✍✜✢✓✔✣✔✄☎✄☎✗✍☞✥✤✦✕✠✡✧☞★✝✠✌✪✩✫☞✍✌✬✕✠✌✮✭✚✌✯✡✰✓✖✎✠✟✰✕✠✡✰✓✖✱✮✲✓✳✟☛✴✵☞✥✤✖✓✖✩✷✶✸✕✹✤✻✺✂✓✳✡✽✼✾✓✿✡❁❀✷✝✠✟☛✺✞❂ ✣✔✟✰✤✻❃✆☞❄✡✻✓✦✤✳✡☛✁✆✟✰✓❅✕✠✌✏✱❇❆❈✓✖✤✻❃✏✕✠✌☎☞✥✩✧✜ ❉❋❊❁●✧❍❏■❑❉▼▲✿◆P❖◗❊❙❘✥❚❱❯ ❲❳❊❩❨✳❬☛❘❭❊❩❪❫❬☛❘✥❴▼❵❛❬☛❘❋◆✂❬✧❜❞❝❢❡❢❪✥❣❤❘❋✐✰❥❫❍❦❣❤❧✿❥♠❣ ♥✷❊✧♦✥♦✥❊☛❥❫♣✰❡✿♦★❣❤❪✥❪❭♦✯qr❧✿♦✍❪★❍s❪❫❡❢❪✥❣❋❬❩❵✂t❳❣✚❥❫♣❢❧❢❬❩❖❦❬☛✉☛❴ ■❢■❢❥❁✈✵❖❏❥❤♦❤▲✇❜①❍s❪❁▲✇❣✚■✻❡ ✐✰❥♠❬☛❪❫❪❋✐☛♣❢❣❤❧❢❚☛❣❤❘ ❲✳❍③②✰❍❏❊⑤④✿♣✿❊❩❧❢✉ ⑥✹❊❁❖❦❬⑧⑦✆❖❦❪✥❬❑⑨✵❣✚♦✍❣✚❊❩❘❭❥❫♣✛◆✂❣❤❧✻❪✥❣❤❘ ⑩✪❣❤❘✥❬❁②✛◆✂❬☛❘✥❝✳❬☛❘❭❊❩❪✍❍❦❬☛❧ ♦✍♣❢❣❤❧❢❚☛❣❤❘❁❶✻❖③❍❷②✻❍❏❊✻✈☎❝✿❊❩❘❭❥❩▲ ②❱❣❤❘✥❬❁②❳▲❸❥♠❬☛❜ ❹✦❺❼❻r❽r❾◗❿❤➀✥❽ ➁❳➂❤➃❸➄✵➅➇➆❫➅❩➈✍➉✲➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉❛➄✵➏❛➂❁➈✏➄❛➐❁➅✚➅❩➋❭➉❛➏✵➋❫➑❳➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✞➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄ ➃❸➌→➆✥➌❞➣✍↔❩↕❦➙➍➛❤➜❏➝❤↕◗➙r➞✵➟☛➙✥➜➡➠✥➢③➤r➙✍➥✹➦✖➝❤➤r➧❤➙✥↕✖➨✹➙✥↕❦➩❢➫❤➜➍➧✆➭❦➯➳➲❁➵✿➸✠➺➡➻✰➼✞➈➾➽➇➉❛➄❛➏ ➉❛➈❄➚♠➃❸➈✍➪✛➏❛➂✚➈✯➊r➂❩➆✥➉r➆❫➊✍➏❛➈❄➉➍➃❸➄❛➏➍➃❸➊✍➄✯➋✥➑✿➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈➾➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✚➄✍➻✹➼✞➈✯➋❭➶❼➓ ➄❛➈❄➉❛➚♠➈✵➏❛➂➇➆❫➏✍➹☛➊✍➋❭➌❤➏❛➉r➆❫➉❛➘❋➏❛➋☎➏❛➂✚➈✯➅❩➋❭➅❁➐❤➒❷➆❫➉✂➚♠➃❸➈✍➪✛➏❛➂➇➆❫➏✹➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈☎➆✥➅❼➓ ➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✚➄➴➌✚➈✍➊✍➈❄➄❛➄r➆❫➉➍➃✇➒❸➘☎➉❛➈✍➷❤➐❤➃❸➉❛➈✠➆✖➽❩➬❤➈❄➎✵➎❁➈r➒❷➆★➘✲➶❩➋♠➐✚➌❁➎☛➹❤➄❛➋❭➔✵➈❳➉❛➈✥➆★➒✇➓ ➏➍➃❸➔✲➈✹➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄➮➆❫➉❛➈❳➔✵➋♠➉❛➈❳➱❩➈❄➬♠➃❸➶❼➒❸➈✂➆✥➌❁➎✵➊★➆❫➌✏➆❫➎➇➆❫➅✚➏✖➏❛➋✹➊✍➐✚➉❛➉❛➈✍➌❤➏ ➌✚➈✍➏❏➪✖➋♠➉❛✃❑➊✍➋❭➌❁➎❤➃❸➏➍➃❸➋♠➌✚➄✍➻❅➼✞➈✞➏❛➂❁➈✍➌❑➅✚➉❛➋♠➅➇➋♠➄❛➈→➆❫➌❑➯➳➲❁➵✿➸❐➆❫➉❛➊r➂❼➃❸➏❛➈✍➊r➓ ➏❛➐✚➉❛➈✪➏❛➂➇➆❫➏☎➄❛➐✚➅❁➅➇➋♠➉❛➏❛➄☎➏➍➪✖➋✞➎❤➃❸➄❛➏➍➃❸➌❁➊✍➏✏✃♠➃❸➌❁➎✚➄☎➋❫➑➮➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈❋➄❛➈✍➉❛➚♠➃❸➊❄➈♠❒ ➛❤❮✚➝❤➜➍➝❤↔❩↕❦➙r➙➡➞✏➄❛➈✍➉❛➚♠➃❸➊✍➈♠➹✂➪❳➂❼➃❸➊r➂⑤➃❸➄❋➏❛➂✚➈❞➏❛➉r➆❫➎❤➃❸➏➍➃❸➋♠➌➇➆★➒✵➑❦➋❭➉❛➔❰➋❫➑➾➉❛➈✥➆★➒✇➓ ➏➍➃❸➔✲➈✹➄❛➈✍➉❛➚♠➃❸➊✍➈✹➎❤➃❸➄❛➊✍➐❁➄❛➄❛➈❄➎✪➃❸➌❋➔✵➋♠➄❛➏✖➋✥➑❢➏❛➂✚➈➴➒✇➃❸➏❛➈✍➉r➆✥➏❛➐❁➉❛➈✲➆❫➌✚➎✆➃❸➌❤➚❭➋❫➒❸➚❭➈✍➄ ➅✚➉❛➈➡➓❏➊✍➋❭➔✵➅❁➐✚➏❛➈✍➎✵➪✦➋♠➉❛➄❛➏➍➓❏➊✥➆❫➄❛➈✿➎✚➈➡➒❷➆✍➘✹➶❩➋♠➐✚➌❁➎✚➄✍➹✚➆✥➌❁➎✹Ï✰➜➍➙r➞❤➢③➤✥↕❦➙➡➞☛➄❛➈✍➉❛➚♠➃❸➊✍➈ ➪❳➂❤➃❸➊➡➂✂➐✚➄❛➈✍➄❢➏❛➂❁➈✿➔✵➈✥➆✥➄❛➐❁➉❛➈❄➎✯➅❩➈❄➉➍➑❦➋♠➉❛➔☎➆✥➌❁➊❄➈➴➋✥➑❩➏❛➂✚➈✿➌❁➈✍➏❏➪✖➋♠➉❛✃❳➃❸➌✯➊❄➋♠➔✯➓ ➅✚➐✚➏➍➃❸➌❁Ð✏➎❁➈➡➒❷➆✍➘❋➶❩➋❭➐❁➌✚➎✚➄✍➻❳➼Ñ➈✂➏❛➂❁➈❄➌❋➅❁➉❛➋❭➅❩➋♠➄❛➈✲➆✯➅➇➆❫➊r✃♠➈✍➏Ò➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð ➔✲➈✍➊➡➂➇➆❫➌❤➃❸➄❛➔✔➏❛➂➇➆❫➏✿➊✥➆✥➌✵➄❛➐❁➅✚➅❩➋❭➉❛➏✖➶❩➋♠➏❛➂✲➋❫➑☛➏❛➂✚➈✍➄❛➈❳➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈Ò➄❛➈✍➉❛➚♠➃❸➊✍➈✍➄ ➆✥➄✯➪✦➈➡➒✇➒❳➆✥➄✵➆✥➊✍➊✍➋❭➔✵➔✵➋❤➎➇➆❫➏❛➈☎➎❩➆✥➏r➆❫Ð❭➉r➆❫➔❐➏❛➉r➆★Ó☎➊❭➻✪➼✞➈☎➆★➒❸➄❛➋✞➎❤➃❸➄❛➊✍➐❁➄❛➄ ➏❏➪✖➋☎➋♠➏❛➂✚➈✍➉✂➆❫➄❛➅➇➈✍➊✍➏❛➄✹➋❫➑Ô➆❫➌✞➋✥➚♠➈✍➉r➆✍➒✇➒✖➯➳➲❁➵✿➸✽➆❫➉❛➊r➂❼➃❸➏❛➈❄➊✍➏❛➐❁➉❛➈❭❒✹➏❛➂❁➈✂➄❛➈✍➉➍➓ ➚♠➃❸➊❄➈P➃❸➌❤➏❛➈✍➉➍➑◗➆❫➊✍➈✲➆❫➌✚➎❋➏❛➂❁➈✂➆✥➎❁➔✂➃❸➄❛➄➍➃❸➋♠➌✞➊✍➋♠➌❤➏❛➉❛➋✥➒❢➊✍➉➍➃❸➏❛➈✍➉➍➃❷➆❤➻ Õ×Ö❸Ø❼❽r❾❦Ù❩Ú➇Û❼➀✥❽❛Ü✇Ù➇Ø ➁❳➂✚➈✯➊❄➐❁➉❛➉❛➈✍➌❤➏✹Ð❭➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌Ý➋✥➑✖➏❛➈r➒❸➈✍➅❁➂✚➋♠➌✚➈☎➌❁➈❄➏➍➪✖➋❭➉❛✃❤➄✂➆❫➌✚➎Ñ➏❛➂✚➈✯➊❄➐❁➉➍➓ ➉❛➈❄➌❤➏✖Ð♠➈✍➌✚➈✍➉r➆✥➏➍➃❸➋♠➌✪➋✥➑✰➊✍➋♠➔✵➅✚➐✚➏❛➈✍➉✖➌✚➈✍➏➍➪✖➋❭➉❛✃❤➄✖➪✖➈✍➉❛➈❳➈✥➆✥➊➡➂☎➎✚➈✍➄➍➃❸Ð♠➌✚➈✍➎✏➏❛➋ ➊★➆❫➉❛➉❛➘✂➄❛➅❩➈✍➊➡➃✇➽➇➊✹➆✥➌❁➎✵➚❭➈✍➉❛➘✹➎❤➃✇Þ☛➈✍➉❛➈✍➌❤➏✖✃♠➃❸➌✚➎✚➄✖➋❫➑☛➏❛➉r➆✍Ó☎➊♠❒❱➆❫➌➇➆★➒❸➋❭Ð✯➚❭➋❫➃❸➊✍➈ ➆✥➌❁➎❋➎❤➃❸Ð❫➃❸➏r➆✍➒❳➎➇➆❫➏r➆❼➻Òß❳➋✥➪✖➈✍➚♠➈❄➉✍➹❩➪✖➃❸➏❛➂❋➏❛➂❁➈✂➎❤➃❸Ð❫➃❸➏➍➃❸à➡➃❸➌✚ÐÑ➋✥➑✿➏❛➈➡➒❸➈✍➅✚➂❁➋❭➌❤➘ ➃❸➌❋➯➳➲❁áÒ➸✛➆❫➌✚➎✪➏❛➂✚➈✠➃❸➌✚➊✍➉❛➈✥➆❫➄➍➃❸➌✚Ð✪➐✚➄❛➈✹➋❫➑✿➔✂➐❤➒❸➏➍➃✇➓➍➔✲➈✍➎❼➃❷➆✵➃❸➌✪➊✍➋❭➔✵➅✚➐❁➏❛➈✍➉ ➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌❩➄✍➹✰➏❛➂❼➃❸➄❳➎❤➃❸➄❛➏➍➃❸➌❁➊❄➏➍➃❸➋♠➌Ñ➃❸➄❳➉r➆❫➅❤➃❸➎❤➒❸➘❋➎❼➃❸➄r➆✥➅❁➅❩➈★➆❫➉➍➃❸➌✚Ð❩➻Òâ✪➈✍➉❛Ð❫➓ ➃❸➌✚Ð✂➏❛➂❁➈✍➄❛➈✦➄❛➋♠➉❛➏❛➄✿➋❫➑☛➄❛➈✍➉❛➚♠➃❸➊❄➈✍➄✳➃❸➌❤➏❛➋✂➆✠➄➍➃❸➌✚Ð❫➒❸➈✠➌✚➈✍➏❏➪✖➋♠➉❛✃❩➹❤➪❳➂❤➃❸➊➡➂☎➪✖➈✦➉❛➈➡➓ ➑s➈✍➉❢➏❛➋✹➂❁➈❄➉❛➈❳➆❫➄✦➆❫➌✲➣✍↔❩↕❦➙➍➛❤➜❏➝❤↕◗➙r➞➴➟❙➙❫➜r➠✥➢③➤r➙★➥❢➦✦➝❤➤r➧❤➙❫↕❩➨➮➙❫↕❦➩❢➫❤➜❏➧➴➭❦➯❛➲✚➵✿➸✠➺➡➹ ➪✦➋♠➐❤➒❸➎✵➘♠➃❸➈➡➒❸➎✏➆✖➄➍➃❸➌❁Ð✥➒❸➈✠➏❛➈➡➒❸➈✍➊❄➋♠➔✵➔✂➐❁➌❤➃❸➊✥➆✥➏➍➃❸➋♠➌❩➄✖➃❸➌❤➑❦➉r➆❫➄❛➏❛➉❛➐✚➊✍➏❛➐✚➉❛➈➴➋✥Þ☛➈✍➉➍➓ ➃❸➌✚Ð✞➆✆➔✂➐❤➒❸➏➍➃❸➏❛➐❁➎✚➈☎➋❫➑Ò➆✥➎❁➚❭➆❫➌❤➏r➆✥Ð♠➈✍➄✍➹✰➃❸➌❁➊➡➒❸➐✚➎❤➃❸➌❩ÐÑ➚❭➆✥➄❛➏✹➈✍➊✍➋♠➌✚➋♠➔✂➃❸➈✍➄✯➋✥➑ ➄❛➊★➆★➒❸➈♠➹✚➐✚➶❼➃❸➷❤➐❤➃❸➏➍➘☎➋❫➑❱➆✥➊✍➊✍➈✍➄❛➄✍➹✚➆✥➌❁➎✂➃❸➔✵➅❁➉❛➋✥➚❭➈✍➎✪➄❛➏r➆✥➏➍➃❸➄❛➏➍➃❸➊✥➆★➒✰➔✂➐❼➒❸➏➍➃❸➅❤➒❸➈✍➬♠➓ ➃❸➌✚Ð❁➻✞➁❳➂✚➈✍➉❛➈✵➃❸➄✵➆✪➶✚➉❛➋❤➆✥➎Ý➊❄➋♠➌✚➄❛➈✍➌❁➄❛➐✚➄✍➹✦➆❫➏✹➒❸➈✥➆❫➄❛➏✂➃❸➌▼➏❛➂❁➈☎➊✍➋❭➔✵➅✚➐❁➏❛➈✍➉ ➌✚➈✍➏❏➪✖➋♠➉❛✃♠➃❸➌✚Ð✪➊✍➋❭➔✵➔✂➐❁➌❤➃❸➏➍➘♠➹☛➏❛➂➇➆❫➏Ò➆✥➌✪➯❛➲✚➵✿➸ã➃❸➄✠➶❩➋❭➏❛➂❋➆✯➪✦➋♠➉❛➏❛➂❤➘❋➆✥➌❁➎ ➆✥➌▼➆✥➊➡➂❤➃❸➈✍➚❭➆❫➶❤➒❸➈❋Ð♠➋ä➆★➒❦➻➾ß➴➋✥➪✖➈✍➚❭➈✍➉✍➹✻➏❛➂✚➈✍➉❛➈☎➆✥➉❛➈✵➔☎➆✥➌❤➘❋➅❩➋❫➒✇➃❸➏➍➃❸➊✥➆✍➒❦➹❳➆✥➎❼➓ ➔✂➃❸➌❼➃❸➄❛➏❛➉r➆✥➏➍➃❸➚♠➈♠➹✦➆❫➌✚➎✞➏❛➈✍➊➡➂✚➌❤➃❸➊✥➆★➒❳➂❤➐✚➉❛➎❤➒❸➈✍➄✯➏❛➋✏➋✥➚♠➈✍➉❛➊✍➋❭➔✵➈✵➶❩➈r➑❦➋♠➉❛➈✂➏❛➂❼➃❸➄ ➚♠➃❸➄➍➃❸➋❭➌Ñ➊★➆❫➌❋➶❩➈✍➊❄➋♠➔✵➈✂➆✯➉❛➈★➆★➒✇➃❸➏➍➘❭➻ å★æ✰ç❛è❦çré➡ê❦ë➍ì✿é➡í✚î✹ï③ð✵ñ✰é➡è❤è❦ò✥ó✥ó♠ô✍ê❦í❦ç❛õ✖ö✍÷✿ø✻ù❢æ✰ú♠ù☎í❦ì❫ê❦ô✍ò✥û✍ì✿ü❢ù✿ý✍ù☎þ❢ê➍érÿ✍í ü❢ù❱þ✁✄✂✆☎✞✝✟ ✠ ö✍÷✹ü✿ý☛✡❋û✍ê❏é➡ÿ✍í❢ü✌☞✧æ✍✂✎✝✏✝✒✑✔✓✒✑✔✝✖✕ ✠ é➡ÿ✥õ✂ö✍÷✹ø✻ù❢æ✰ú♠ù→é➡ÿ✥õ✂ü✿ý✖✡ í❦ì✥ê❦ô✍ò✥û✍ì✗☞☛ô✥ô✍ó♠ç❛ê❏é➡í✆✘ ✙rç❢ù❢û✍ê❦ç❛ç✔✚✖ç❛ÿ✍í☛ü✌☞☛æ✛✂✜✝✟✢✣✑✞✢☛✤✞✥☛✝✻ñ✦✘ í❦ì✹í❦ì❫ç✧☞✧ô❄ê◗ó❭ô✍ê❏é➡í✆✘ ô✍ÿ ★ô❄ê✻ü✿é➡í✆✘ ô❄ÿ♠é✞✩❤æ✻ç❛è❦çré➡ê❦ë➍ì❳ï③ÿ✟✘ í✆✘ é➡í✆✘ ✙rç❛è✄✪ ✫✠➌❁➈☎➋❫➑Ô➏❛➂❁➈☎➔✵➋❭➄❛➏✯➚❭➈✍➬♠➃❸➌❁Ð✞➏❛➈✍➊r➂❁➌❤➃❸➊✥➆✍➒✠➅✚➉❛➋♠➶❤➒❸➈✍➔✵➄☎➏❛➂➇➆❫➏✂➶❼➒❸➋❤➊r✃❤➄ ➏❛➂✚➈✪➅➇➆❫➏❛➂❞➏❛➋✥➪❳➆❫➉❛➎✚➄☎➆❫➌❞➯➳➲❁➵✿➸ ➃❸➄☎➏❛➂❩➆✥➏✵➋❫➑Ò➄❛➐✚➅❁➅❩➋❭➉❛➏➍➃❸➌❁Ð❞➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈ ➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄Ñ➃❸➌✫➆▼➅❩➆✥➊➡✃❭➈✍➏❋➌✚➈✍➏➍➪✖➋❭➉❛✃❩➻✭✬✖➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈→➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✚➄ ➆❫➉❛➈✹➷❤➐❤➃❸➏❛➈✵➎❤➃✇Þ☛➈✍➉❛➈✍➌❤➏❳➑❦➉❛➋♠➔✾➄❛➏r➆✥➌❁➎➇➆❫➉❛➎✞➎➇➆❫➏r➆✵➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄✍➹✿➆✥➌❁➎✏➉❛➈➡➓ ➷❤➐❤➃❸➉❛➈✖➄❛➈✍➉❛➚♠➃❸➊✍➈✖➏❛➂➇➆❫➏❢➊★➆❫➌✚➌❁➋❭➏❢➶❩➈✳➎✚➈➡➒✇➃❸➚❭➈✍➉❛➈✍➎✵➪✖➃❸➏❛➂❤➃❸➌✵➏❛➂❁➈✿➏➍➘❤➅❤➃❸➊✥➆✍➒❩➎➇➆❫➏r➆ ➄❛➈✍➉❛➚♠➃❸➊✍➈✏➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈♠➻❋➯❏➌▼➲❤➈✍➊✍➏➍➃❸➋❭➌✯✮☎➪✖➈✵➎❤➃❸➄❛➊✍➐✚➄❛➄✵➏❛➂❁➈✲➌❩➆✥➏❛➐❁➉❛➈☎➋❫➑ ➉❛➈✥➆✍➒✇➓➍➏➍➃❸➔✲➈▼➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄✞➆❫➏✵➒❸➈✍➌✚Ð♠➏❛➂✍✰✹➂✚➈✍➉❛➈♠➹❳➂✚➋✥➪✖➈✍➚♠➈❄➉✍➹✳➃❸➏✪➄❛➐❤Ó☎➊✍➈❄➄ ➏❛➋▼➋❭➶❁➄❛➈✍➉❛➚❭➈▼➏❛➂➇➆❫➏☎➋♠➌✚➈Ñ➄r➆✍➒✇➃❸➈✍➌❤➏❋➊r➂❩➆✥➉r➆❫➊✍➏❛➈✍➉➍➃❸➄❛➏➍➃❸➊⑧➋❫➑✂➏❛➂✚➈Ñ➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈ ➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄Ñ➪✦➈❋➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉✪➃❸➄✪➏❛➂➇➆❫➏☎➏❛➂❁➈❄➘✸➉❛➈✍➷❤➐❤➃❸➉❛➈❞➆Ñ➶❩➋❭➐❁➌✚➎✫➋♠➌ ➏❛➂✚➈✆➎✚➈➡➒✇➃❸➚❭➈✍➉❛➘▼➎✚➈➡➒❷➆✍➘▼➋❫➑❳➈★➆❫➊r➂▼➅❩➆✥➊➡✃❭➈✍➏✔✱❤➻☎➼✸➂❤➃✇➒❸➈❋➏❛➂❼➃❸➄✂➶❩➋❭➐❁➌✚➎❞➔☎➆✍➘ ➶❩➈☎➄❛➏r➆❫➏➍➃❸➄❛➏➍➃❸➊★➆★➒❦➹✖➃❸➌❞➏❛➂❁➈☎➄❛➈✍➌✚➄❛➈☎➏❛➂❩➆✥➏✯➄❛➋❭➔✵➈☎➄❛➔☎➆★➒✇➒❳➑❦➉r➆✥➊✍➏➍➃❸➋♠➌⑧➋❫➑Ò➏❛➂❁➈ ➅➇➆❫➊r✃♠➈✍➏❛➄✵➔➾➆★➘✞➑❏➆★➃✇➒❳➏❛➋✞➆✥➉❛➉➍➃❸➚♠➈❋➶❤➘✞➏❛➂❼➃❸➄☎➶❩➋❭➐❁➌✚➎☛➹✖➏❛➂✚➈✆➶➇➋♠➐✚➌❁➎❞➃❸➏❛➄❛➈➡➒✇➑ ➔✂➐❁➄❛➏❳➶❩➈Ò✃❤➌❁➋✥➪❳➌▼➝✠Ï✰➜r➢③➫❤➜➡➢❸➻✻➁❳➂✚➈✹➏❛➉r➆❫➎❤➃❸➏➍➃❸➋♠➌➇➆★➒✿➎➇➆❫➏r➆✯➄❛➈✍➉❛➚♠➃❸➊❄➈✂➆❫➉❛➊r➂❼➃✇➓ ➏❛➈✍➊✍➏❛➐✚➉❛➈✪➐✚➌✚➎❁➈✍➉➍➒❸➘♠➃❸➌✚Ð❞➊✍➋♠➔✲➅❁➐✚➏❛➈✍➉☎➌❁➈❄➏➍➪✖➋❭➉❛✃❤➄✯➂➇➆❫➄✵➌✚➋❋➑❏➆❫➊➡➃✇➒✇➃❸➏➍➃❸➈❄➄✆➑❦➋❭➉ ➅✚➉❛➈✍➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❑➉❛➈✍➄❛➋❭➐❁➉❛➊✍➈❄➄✪➋♠➉✲➎❁➈✍➌❤➘♠➃❸➌✚Ð✸➄❛➈❄➉❛➚♠➃❸➊✍➈Ñ➐✚➅❩➋❭➌❑➋❫➚❭➈✍➉➍➒❸➋❤➆✥➎✧➹ ➆❫➌✚➎❋➏❛➂❤➐✚➄Ô➃❸➄Ò➐❁➌➇➆❫➶❤➒❸➈✵➏❛➋✂➔✵➈✍➈✍➏❳➏❛➂❤➃❸➄✠➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈✯➉❛➈❄➷❤➐❼➃❸➉❛➈✍➔✲➈✍➌❤➏✍➻ ➁➴➂✚➈✍➉❛➈➡➑s➋♠➉❛➈♠➹❳➃❸➌⑤➋❭➉❛➎❁➈❄➉❋➏❛➋Ý➂➇➆❫➌✚➎❤➒❸➈▼➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈❞➏❛➉r➆✍Ó☎➊♠➹➮➆❫➌❑➈✍➌❤➓ ➂➇➆❫➌✚➊✍➈✍➎❋➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈✠➃❸➄✖➌✚➈✍➈✍➎✚➈✍➎✆➑s➋♠➉✦➆❫➌✪➯➳➲❁➵✿➸P➻ä➼Ñ➈❳➃❸➎❁➈✍➌❤➏➍➃✇➑s➘➾➑s➋♠➐✚➉ ✃♠➈❄➘❋➊✍➋♠➔✵➅➇➋♠➌✚➈✍➌❤➏❛➄✯➏❛➋☎➏❛➂❤➃❸➄✂➆❫➉❛➊r➂❼➃❸➏❛➈❄➊✍➏❛➐❁➉❛➈❭➻✵➁❳➂❁➈✹➽➇➉❛➄❛➏✠➅❤➃❸➈✍➊✍➈☎➋❫➑✦➏❛➂❁➈ ➆❫➉❛➊r➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈✯➃❸➄❳➏❛➂❁➈✹➌➇➆❫➏❛➐✚➉❛➈✯➋✥➑✿➏❛➂❁➈✹➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏❛➄✠➔☎➆✥➎❁➈✂➶❤➘☎➏❛➂❁➈ ➌✚➈✍➏➍➪✖➋❭➉❛✃✵➪❳➂❁➈❄➌➾➃❸➏❳➅✚➉❛➋♠➔✯➃❸➄❛➈❄➄➴➏❛➋✂➎✚➈➡➒✇➃❸➚♠➈✍➉➮➆P➊❄➈✍➉❛➏r➆★➃❸➌✏➷❤➐❩➆✍➒✇➃❸➏➍➘✪➋✥➑✰➄❛➈✍➉➍➓ ➚♠➃❸➊✍➈♠➻✿➼✞➈➴➃❸➎✚➈✍➌❤➏➍➃✇➑❦➘☎➏➍➪✦➋✹➄❛➋♠➉❛➏❛➄✖➋✥➑✻➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏❛➄✍➹➇➛❼❮✚➝❤➜➍➝❤↔➇↕◗➙r➙r➞✳➆✥➌❁➎ Ï✰➜➍➙r➞❤➢③➤✥↕❦➙➡➞♠➻ ➵ ➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❋➄❛➈❄➉❛➚♠➃❸➊✍➈✯➃❸➄✹➆✂➔☎➆✖✲➳➋♠➉✹➆✥➄❛➅❩➈✍➊✍➏Ò➋❫➑✿➋❭➐❁➉Ò➅❩➆✥➅❩➈✍➉✍➻ ➼✸➂❤➃✇➒❸➈✵➏❛➂✚➈✠➃❸➎❁➈★➆➾➋✥➑✿➅❁➉❛➈❄➎❼➃❸➊✍➏❛➈❄➎Ñ➄❛➈✍➉❛➚♠➃❸➊❄➈✯➂➇➆❫➄❳➶❩➈✍➈❄➌❋➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉❛➈✍➎Ñ➶➇➈➡➓ ➑❦➋❭➉❛➈♠➹❤➏❛➂✚➈➴➃❸➄❛➄❛➐✚➈✍➄❳➏❛➂❩➆✥➏✖➄❛➐❁➉❛➉❛➋❭➐❁➌✚➎✆➃❸➏✦➂❩➆✍➚♠➈Ò➌❁➋❭➏✍➹❤➏❛➋✂➋♠➐✚➉✖✃❤➌❁➋✥➪✖➒❸➈✍➎✚Ð♠➈♠➹ ➶❩➈✍➈❄➌❋➊✥➆❫➉❛➈➡➑s➐❼➒✇➒❸➘❋➈✍➬❤➅❤➒❸➋♠➉❛➈❄➎✧➻ ➁➴➂✚➈✹➄❛➈✍➊✍➋❭➌❁➎❋➅❤➃❸➈✍➊✍➈✹➋✥➑✿➏❛➂❁➈✂➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈✹➃❸➄✠➏❛➂✚➈✹➄❛➈✍➉❛➚♠➃❸➊✍➈✹➃❸➌❤➏❛➈✍➉➍➓ ➑❏➆✥➊✍➈♠➹✻➃❦➻ ➈♠➻❸➹❢➏❛➂❁➈☎➄❛➈✍➏✹➋✥➑Ò➅❩➆✥➉r➆❫➔✵➈✍➏❛➈❄➉❛➄✯➅➇➆❫➄❛➄❛➈✍➎❞➶❩➈❄➏➍➪✖➈✍➈❄➌▼➏❛➂❁➈✲➄❛➋♠➐✚➉❛➊✍➈ ➆❫➌✚➎❞➏❛➂❁➈☎➌✚➈✍➏➍➪✦➋♠➉❛✃❩➻❋➁❳➂✚➈✵➄❛➈✍➉❛➚♠➃❸➊✍➈☎➃❸➌❤➏❛➈✍➉➍➑❏➆❫➊❄➈✪➔✂➐❁➄❛➏✹➃❸➌❁➊r➒❸➐❁➎✚➈❋➶❩➋♠➏❛➂ ➏❛➂✚➈✖➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌☎➋❫➑➇➏❛➂✚➈✖➷❤➐❩➆✍➒✇➃❸➏➍➘✵➋✥➑❩➄❛➈❄➉❛➚♠➃❸➊✍➈❳➏❛➂❁➈✖➌✚➈✍➏❏➪✖➋♠➉❛✃✹➪✖➃✇➒✇➒ ➎✚➈➡➒✇➃❸➚♠➈✍➉✍➹❤➑❦➐❤➒✇➽✚➒✇➒✇➃❸➌❁Ð✵➏❛➂✚➈✖➌✚➈✍➈✍➎✵➋✥➑☛➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌❩➄✖➏❛➋Ò✃❤➌❁➋✥➪❞➪❳➂❁➈❄➌✯➏❛➂✚➈➡➃❸➉ ➅➇➆❫➊r✃♠➈✍➏❛➄✹➪✖➃✇➒✇➒✦➆❫➉❛➉➍➃❸➚❭➈♠➹✰➆❫➌✚➎✞➏❛➂❁➈✂➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌▼➋✥➑✖➏❛➂✚➈✂➄❛➋♠➐✚➉❛➊✍➈✒✳ ➄ ➏❛➉r➆★Ó➾➊♠➹✿➏❛➂❁➈✍➉❛➈❄➶❤➘✸➆✍➒✇➒❸➋✥➪Ô➃❸➌✚Ð❞➏❛➂❁➈✏➌❁➈❄➏➍➪✖➋❭➉❛✃Ñ➏❛➋✞✃❤➌❁➋✥➪✖➒❸➈✍➎✚Ð♠➈✥➆✥➶❼➒❸➘✫➆✍➒✇➓ ➒❸➋❤➊✥➆✥➏❛➈✪➉❛➈✍➄❛➋❭➐❁➉❛➊✍➈❄➄✍➻❋➯➍➌✞➏❛➂❼➃❸➄✲➅❩➆✥➅❩➈✍➉✵➪✖➈➾➆❫➏❛➏❛➈✍➔✵➅✚➏✂➏❛➋✪➃❸➎❁➈✍➌❤➏➍➃✇➑s➘▼➏❛➂❁➈ ➊✍➉➍➃❸➏➍➃❸➊✥➆✍➒✰➆❫➄❛➅➇➈✍➊✍➏❛➄✿➋❫➑☛➏❛➂✚➈❳➄❛➈✍➉❛➚♠➃❸➊✍➈✖➃❸➌❤➏❛➈✍➉➍➑❏➆✥➊✍➈♠➹✚➆✥➌❁➎✵➋✥Þ☛➈✍➉✦➆➴➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉ ➃❸➌❤➏❛➈✍➉➍➑❏➆✥➊✍➈✂➆❫➄Ô➆❫➌☎➈✍➬✚➆❫➔✲➅❼➒❸➈♠➻✦➼Ñ➈➮➆❫➎✚➎❁➉❛➈❄➄❛➄Ô➃❸➌✏➅❩➆✥➄❛➄➍➃❸➌❁Ð☎➏❛➂✚➈✠➌✚➈✍➈✍➎✵➑❦➋❭➉ ➈✍➌❤➑❦➋♠➉❛➊❄➈✍➔✵➈✍➌❤➏Ò➋❫➑❢➏❛➂✚➈✍➄❛➈✂➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌✚➄✍➻ ➁➴➂✚➈❳➏❛➂❼➃❸➉❛➎✏➅❼➃❸➈❄➊✍➈✠➋✥➑❢➏❛➂✚➈✠➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈✠➃❸➄✖➏❛➂✚➈✠➅➇➆❫➊r✃♠➈✍➏✦➄❛➊➡➂✚➈✍➎✚➐❼➒✇➓ ➃❸➌✚Ð✪➶❩➈✍➂➇➆✍➚♠➃❸➋♠➉✂➋❫➑✿➌✚➈✍➏➍➪✦➋♠➉❛✃✪➄❛➪✖➃❸➏❛➊r➂❁➈❄➄✹➌❁➈✍➈❄➎❁➈✍➎✞➏❛➋✵➔✵➈❄➈✍➏✠➏❛➂✚➈✍➄❛➈✂➄❛➈✍➉➍➓ ➚♠➃❸➊✍➈✏➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏❛➄✍➻❋➼✞➈☎➎❼➃❸➄❛➊✍➐✚➄❛➄✵➶❩➋❭➏❛➂▼➏❛➂✚➈✪➆✥➊✍➏❛➐➇➆★➒❳➄❛➊r➂❁➈❄➎❁➐❤➒✇➃❸➌❁Ð ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄❳➏❛➋✠➶❩➈❳➐✚➄❛➈✍➎✯➃❸➌✵➏❛➂✚➈❳➄❛➪✖➃❸➏❛➊➡➂✚➈✍➄✍➹✚➆✥➄✳➪✖➈r➒✇➒☛➆✥➄✳➏❛➂✚➈❳➄❛➊➡➂✚➈✍➎✚➐❼➒✇➓ ➃❸➌✚Ð✠➃❸➌❼➑❦➋❭➉❛➔☎➆❫➏➍➃❸➋❭➌✵➏❛➂➇➆❫➏❢➔✂➐❁➄❛➏❢➶❩➈✿➊✥➆✥➉❛➉➍➃❸➈✍➎✯➃❸➌✵➅➇➆❫➊r✃♠➈✍➏❢➂✚➈✥➆✥➎❁➈✍➉❛➄✍➻❱➁➴➂❤➃❸➄ ✱ ý☛✘ ÿ✥ë❛ç❢í❦ì✥ç✳í❦ç❛ê✆✚✵✴✷✶☛✸✟✹✻✺✼✘ è✻í❦ô✍è❦è❦ç❛õ✠érê◗ô❄ò❫ÿ✥õ❳ñ✽✘ í❦ì✠û✍ê❦çré➡í☛érö♠érÿ❫õ✥ô✍ÿ✾✘ÿ✠í❦ì✥ç ê❦ç❛è❦í✖ô★ í❦ì❫çÒó❭é➡ó♠ç❛ê ✠ ñ☛ç❳ÿ✥ç❛ç❛õ✵í❦ô✿✘õ✥ç❛ÿ✍í✆✘ ★÷☎è❦ç✔✙➡ç❛ê❏é❀✩❩õ✟✘❁✚ç❛ê◗ç❛ÿ❄í❂✚✦çré➡ÿ✟✘ ÿ✥û✍è✖í❦ô í❦ì✥ç✠í❦ç❛ê✆✚❃✪✹ù❢ÿ❅❄❃❆✻❇✔❈✎✶☛❇✔❈✚ö♠ô✍ò✥ÿ✥õ✵ô✍ÿ✲õ❫ç✔✩ é÷❉✘ èÔéPè❦í❏é➡í❦ç✔✚✖ç❛ÿ✍í✿í❦ì♠érí✖ÿ✥ô✍ÿ✥ç➴ô★ í❦ì✥ç ★ò✥í❦ò✥ê◗ç✿õ✥ç✔✩✇é❛÷✍è✻ñ✦✘ ✩ ✩❁ç✔❊✍ë❛ç❛ç❛õ❳í❦ì❭é➡í✰é❀✚✖ô✍ò✥ÿ✍í✔✪➇ù❋❆✒✶☛●✷❍❏■✞❄✟❑✄❍✆✶✰ö♠ô✍ò✥ÿ✥õ❃✘ è✻í❦ì✥ç ✚❳é❀❊✏✘✚❳é❀✩✍✙ré❀✩ò✥ç✠ô★ é➴è❦ç❛í✿ô★ ô✍ö✥è❦ç❛ê▲✙rç❛õ✵õ✥ç✔✩✇é❛÷✍è✔✪Òý✍í❏é➡í✆✘ è❦í✆✘ ëré❀✩☛ö♠ô❄ò❫ÿ✥õ✥è❳é❀✩ ✩ ôrñ ★ô✍ê✿é❳ë❛ç❛ê◗í❏é❀✘ÿ✹ó♠ç❛ê❦ë❛ç❛ÿ✍í❏é➡û✍ç❢ô★ ✙☛✘ ô✏✩✇é➡í✆✘ ô❄ÿ❫è✿ô★ í❦ì✥ç✖ö❭ô✍ò✥ÿ❫õ✻▼➇é➡ö✥è❦ô☛✩ ò✥í❦ç✖ö❭ô✍ò✥ÿ❫õ✥è é❀✩ ✩ ôrñ❋ÿ❫ô❄ÿ❫ç☛✪
part of the architecture must be carefully considered; since it Parekh and 1 allager ([-4, 27]), 2acobson and Floyd([-4]) lust be executed for every packet it must not be so complex and ' azar, Hyman, and Pacifici([-2,-8]) have all cor as to effect tribu ted The final part of the architecture is the means by which Finally, in Section-2, we conclude our paper wi the traffic and service commit ments get est ablished. 2 learl view of our results and a brief discussion of related economic the ability of the network to meet its service commitments is issues. The s ppendix contains details relating to the simu elated to the criteria the net work uses to decide whether to lation results that are presented in Sections 5-7 accept another request for service. While we do not present a specific algorit hm to regulate the admission of new sources, we show the re lation between the other parts of our proposal o Pro u drtids of RdalaTip d Trar c and a general approach to the ad mission control problem oil A lass of RdalaTip d Auu li directl lated to the nature of real-time traffic. for inst ance, the In the discu on a particular class sues of routing, datagram congestion control, and the inter of real-time applications which we dub paac ack applica action of administrative domains all pose interesting chal tions. In a play-back application, the source takes sor any final architectural proposal for an ISPN must solve thes he net work ine vit ably introduces some variation in the d longs anding problems. It is import lay of each delivered packet. This variation has traditionally t hat we do not believe that the architect ural choices we ad- been called jitter. The receiver de packetizes the data and vocate here for real-time traffic unnecessarily rest rict the then attempts to faithfully play back the sig nal. This is done scope of solutions to these ot her problems by buffering the incoming data to remove the network in 2 we begin with a disc ussion of the nature of realtime traffic. &ced itter and then replaying the signal at some This paper has -2 Sections and an s ppendix. In Section ted In particular, we note that some real-time applications can pac- ack point can be used to reconstruct io p 37c;ack point. 5 ny dat a that arrives before dat adapt te ent net work condit ions. This leads us to pi arriving after the p3c- ack point is useless in re cting pose, in Section 8, that an ISPN support two kinds of rea the real-time signal. For the purposes of this ve as- ime service commitments: guaranteed service and predicted sume that all such application sufficient store all packets which arrive before the play-back point; we ing algorithm which is a nonuniformly weig hted ve the Igorithm discussed in Reference [4], and Not all real-time applications are p3ac- ack application then refer to a recent result due to Parekh and I allager( ght imagine a visualization apl References [-4, 27]) which states that, under certain condi- which merely displayed the image encoded in each packet of arbitrary topology. We then turn, in Sections 5 and< ions, this algorit hm delivers guaranteed se enever it arrived). Howe ver, we believe the vast majority lications, including most video and audio applications, will fit this paradigm. Furthermore, non dicted service. The predicted service scheduling algorithm play-back applications can still use the real-time network wel ideas: that of ervice provided b ct ure, although th time context, and that of correlating the qu dela ot specifi tailored to their needs packet at successive nodes in th to reduce del Play-back real-time applications have se veral service quirements that inform our design proposal. First, sinc 7, presenting a unified scheduling algorithm which provide t here is often real-time interaction bet ween the two end oth guaranteed and predicted service. I iven the current of an application, as in a voice conversation, the application frenzy of activity in the design of real-time scheduling algo- performance is sensitive to the data delivery delay; in general ithms, we do not expect that the algorithm presented here lower delay is much preferable. Second, in order to set the will be the final word on the matter; howe ver, we do hope lay-back point, the application needs to have some info that the at embodied therein will be of lasting value mation(pre ferably an absolute or statistical bound)about In rticular, we think that the e insight underlying the delays that each packet will experience. Third, since all sign, that it is necessary to distinguish between the two basic dat a is buffered until the play-back point, the application is rinciples of isolation and sharing, is both fundamental and indifferent as to when data is delivered as long as it arrive before the play-back point. This turns out to be a crucial ince the service interface will be invoked by applications, aktime service interface will outlive which are. Fourth, these pla particular underlying network mechanism. Thus, we have tolerate the loss of a certain fraction of packe ts with only a pted in our proposal face which is minimal distortion in the sig nal. Therefore, the play-back enough to a scussed brief point need not be so delayed that absolutely every packet arrives before hand covered in Section-7. In order to build up sufficient context to meaningfully compare our work The delay in the network derives from se veral causes. There ay the detailed discussion of related work until Section is in practice a large fixed component to the delay, caused However, we wish to note here that our work borrows heay ly from the rapidly growing literature on pro viding reah pus that the rece上h time service in packet networks. In particular, the works of
➅➇➆❫➉❛➏❢➋✥➑❩➏❛➂✚➈❳➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈✖➔✂➐❁➄❛➏❢➶❩➈✦➊✥➆❫➉❛➈➡➑s➐❼➒✇➒❸➘✂➊✍➋♠➌✚➄➍➃❸➎✚➈✍➉❛➈✍➎✍✰❁➄➍➃❸➌✚➊✍➈✖➃❸➏ ➔✂➐✚➄❛➏✳➶❩➈✦➈✍➬❤➈✍➊✍➐✚➏❛➈✍➎✯➑❦➋❭➉✳➈❄➚♠➈✍➉❛➘✹➅➇➆❫➊r✃♠➈✍➏❢➃❸➏✿➔✯➐✚➄❛➏❢➌❁➋❭➏✳➶➇➈✖➄❛➋✹➊✍➋♠➔✲➅❼➒❸➈✍➬ ➆✥➄❳➏❛➋✵➈➡Þ☛➈✍➊❄➏✠➋✥➚♠➈✍➉r➆✍➒✇➒❢➌✚➈✍➏➍➪✦➋♠➉❛✃✪➅➇➈✍➉➍➑❦➋♠➉❛➔➾➆❫➌✚➊✍➈♠➻ ➁❳➂❁➈✠➽➇➌➇➆★➒❢➅➇➆❫➉❛➏❳➋✥➑✿➏❛➂❁➈➮➆❫➉❛➊r➂❼➃❸➏❛➈✍➊❄➏❛➐❁➉❛➈✯➃❸➄❳➏❛➂✚➈✹➔✵➈✥➆✥➌❁➄❳➶❤➘☎➪❳➂❼➃❸➊r➂ ➏❛➂✚➈✖➏❛➉r➆✍Ó☎➊✖➆✥➌❁➎✂➄❛➈✍➉❛➚♠➃❸➊✍➈❳➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈❄➌❤➏❛➄❢Ð♠➈✍➏❢➈✍➄❛➏r➆✥➶❼➒✇➃❸➄❛➂✚➈✍➎☛➻✁✳➒❸➈✥➆✥➉➍➒❸➘♠➹ ➏❛➂✚➈❳➆✥➶❼➃✇➒✇➃❸➏❏➘✆➋✥➑☛➏❛➂✚➈✖➌❁➈❄➏➍➪✖➋❭➉❛✃✹➏❛➋✹➔✵➈✍➈✍➏✻➃❸➏❛➄❢➄❛➈✍➉❛➚♠➃❸➊❄➈➴➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏❛➄✳➃❸➄ ➉❛➈r➒❷➆❫➏❛➈✍➎✏➏❛➋✹➏❛➂❁➈❳➊✍➉➍➃❸➏❛➈✍➉➍➃❷➆✂➏❛➂❁➈❳➌✚➈✍➏➍➪✖➋❭➉❛✃✵➐✚➄❛➈✍➄✖➏❛➋✹➎✚➈✍➊➡➃❸➎✚➈✹➪➴➂✚➈✍➏❛➂✚➈✍➉✿➏❛➋ ➆✥➊✍➊✍➈✍➅✚➏Ò➆❫➌✚➋♠➏❛➂✚➈✍➉❳➉❛➈✍➷❤➐✚➈✍➄❛➏✳➑❦➋❭➉✖➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻✦➼✸➂❤➃✇➒❸➈✯➪✦➈✠➎✚➋✯➌✚➋♠➏❳➅✚➉❛➈✍➄❛➈✍➌❤➏ ➆✖➄❛➅❩➈❄➊➡➃✇➽❩➊Ò➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✽➏❛➋❳➉❛➈✍Ð♠➐❤➒❷➆❫➏❛➈✖➏❛➂✚➈✦➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌☎➋❫➑➇➌✚➈✍➪▼➄❛➋♠➐✚➉❛➊✍➈✍➄✍➹ ➪✦➈✖➄❛➂❁➋✥➪❞➏❛➂✚➈✖➉❛➈➡➒❷➆❫➏➍➃❸➋❭➌☎➶❩➈✍➏➍➪✦➈✍➈✍➌✂➏❛➂❁➈✖➋❭➏❛➂❁➈❄➉✳➅➇➆❫➉❛➏❛➄❢➋✥➑☛➋♠➐✚➉❢➅❁➉❛➋❭➅❩➋♠➄r➆✍➒ ➆✥➌❁➎✞➆✂Ð♠➈✍➌✚➈✍➉r➆★➒ ➆❫➅✚➅❁➉❛➋ä➆❫➊r➂Ñ➏❛➋✲➏❛➂❁➈✂➆✥➎❁➔✯➃❸➄❛➄➍➃❸➋❭➌Ñ➊❄➋♠➌❤➏❛➉❛➋❫➒❱➅❁➉❛➋❭➶❼➒❸➈❄➔✪➻ ➁❳➂❁➈❄➉❛➈✂➆❫➉❛➈✂➆★➒❸➄❛➋☎➔☎➆✥➌❤➘✏➆❫➉❛➊r➂❼➃❸➏❛➈✍➊❄➏❛➐❁➉r➆✍➒❢➃❸➄❛➄❛➐❁➈✍➄Ò➌❁➋❭➏❳➎❼➃❸➉❛➈✍➊❄➏➍➒❸➘❋➉❛➈➡➓ ➒❷➆✥➏❛➈✍➎✞➏❛➋✪➏❛➂✚➈✯➌➇➆❫➏❛➐✚➉❛➈✵➋❫➑✦➉❛➈✥➆✍➒✇➓➍➏➍➃❸➔✲➈✆➏❛➉r➆✍Ó☎➊✒✰❩➑❦➋❭➉➴➃❸➌✚➄❛➏r➆❫➌✚➊✍➈♠➹❢➏❛➂✚➈✯➃❸➄➍➓ ➄❛➐✚➈✍➄❳➋❫➑❱➉❛➋♠➐✚➏➍➃❸➌❁Ð✚➹❩➎➇➆❫➏r➆✥Ð♠➉r➆❫➔✾➊❄➋♠➌✚Ð♠➈✍➄❛➏➍➃❸➋❭➌❋➊✍➋♠➌❤➏❛➉❛➋✥➒❦➹✰➆❫➌✚➎✆➏❛➂✚➈✠➃❸➌❤➏❛➈✍➉➍➓ ➆✥➊✍➏➍➃❸➋♠➌→➋❫➑Ò➆✥➎❁➔✂➃❸➌❼➃❸➄❛➏❛➉r➆✥➏➍➃③➚❭➈Ñ➎✚➋♠➔☎➆✍➃❸➌❁➄✏➆✍➒✇➒❳➅❩➋♠➄❛➈☎➃❸➌❤➏❛➈✍➉❛➈✍➄❛➏➍➃❸➌✚Ð▼➊r➂❩➆✍➒✇➓ ➒❸➈❄➌❁Ð❭➈✍➄✍➻▼➼✞➈☎➎❁➋❋➌✚➋♠➏☎➆✥➎❁➎✚➉❛➈✍➄❛➄✵➏❛➂✚➈✍➄❛➈✵➃❸➄❛➄❛➐✚➈✍➄✯➃❸➌❞➏❛➂❼➃❸➄✲➅❩➆✥➅❩➈✍➉✍➹❳➆✥➌❁➎ ➆✥➌❤➘✠➽❩➌➇➆★➒❙➆❫➉❛➊r➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉r➆★➒☛➅✚➉❛➋♠➅❩➋❭➄r➆★➒✚➑❦➋♠➉❱➆❫➌✂➯❛➲✚➵✳➸❞➔✯➐✚➄❛➏✻➄❛➋✥➒❸➚♠➈❳➏❛➂✚➈✍➄❛➈ ➒❸➋❭➌❁Ð❭➄❛➏r➆❫➌✚➎❼➃❸➌✚Ð❑➅✚➉❛➋♠➶❤➒❸➈✍➔✵➄✍➻✫➯➍➏✯➃❸➄✵➃❸➔✵➅➇➋♠➉❛➏r➆❫➌❤➏☎➏❛➋▼➌✚➋♠➏❛➈❭➹✖➂✚➋❫➪✦➈✍➚♠➈✍➉✍➹ ➏❛➂➇➆❫➏❳➪✖➈Ò➎❁➋✂➌✚➋♠➏❳➶❩➈➡➒✇➃❸➈✍➚❭➈✵➏❛➂➇➆❫➏✖➏❛➂✚➈✂➆❫➉❛➊r➂❼➃❸➏❛➈❄➊✍➏❛➐❁➉r➆✍➒❢➊r➂❁➋✥➃❸➊✍➈✍➄Ò➪✖➈✹➆✥➎❼➓ ➚❭➋❤➊✥➆❫➏❛➈✞➂✚➈✍➉❛➈Ñ➑❦➋❭➉✆➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈▼➏❛➉r➆✍Ó☎➊Ñ➐✚➌✚➌❁➈✍➊❄➈✍➄❛➄r➆❫➉➍➃✇➒❸➘✛➉❛➈❄➄❛➏❛➉➍➃❸➊✍➏✪➏❛➂✚➈ ➄❛➊❄➋♠➅❩➈✂➋❫➑✿➄❛➋✥➒❸➐❁➏➍➃❸➋❭➌❁➄✹➏❛➋✵➏❛➂✚➈✍➄❛➈✹➋❭➏❛➂❁➈✍➉Ò➅❁➉❛➋❭➶❼➒❸➈❄➔✵➄✍➻ ➁❳➂❼➃❸➄✦➅❩➆✥➅❩➈✍➉❳➂❩➆✥➄✄✂✖✮✯➲❤➈✍➊✍➏➍➃❸➋❭➌❁➄Ò➆✥➌❁➎✏➆❫➌✆☎Ò➅❁➅❩➈❄➌❁➎❤➃❸➬❩➻✖➯❏➌✪➲❤➈✍➊✍➏➍➃❸➋❭➌ ✮✖➪✖➈✿➶❩➈✍Ð✥➃❸➌✵➪✖➃❸➏❛➂✵➆✖➎❤➃❸➄❛➊✍➐✚➄❛➄➍➃❸➋♠➌☎➋❫➑➇➏❛➂❁➈✿➌➇➆❫➏❛➐✚➉❛➈✖➋❫➑➇➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈❳➏❛➉r➆✍Ó☎➊♠➻ ➯❏➌❋➅❩➆✥➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✍➹❢➪✖➈✹➌✚➋♠➏❛➈✂➏❛➂➇➆❫➏✠➄❛➋❭➔✵➈✂➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈☎➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✚➄✵➊✥➆❫➌ ➆✥➎❩➆✥➅❁➏✹➏❛➋☎➊✍➐✚➉❛➉❛➈✍➌❤➏✹➌✚➈✍➏➍➪✦➋♠➉❛✃❋➊✍➋❭➌❁➎❤➃❸➏➍➃❸➋♠➌✚➄✍➻✵➁❳➂❤➃❸➄✠➒❸➈✥➆❫➎✚➄✹➐❁➄✹➏❛➋☎➅✚➉❛➋❫➓ ➅➇➋♠➄❛➈♠➹❩➃❸➌✞➲❤➈✍➊✍➏➍➃❸➋❭➌✞✝❼➹❩➏❛➂➇➆❫➏➮➆❫➌❋➯➳➲❁➵✿➸⑤➄❛➐✚➅❁➅❩➋❭➉❛➏✠➏❏➪✖➋✵✃♠➃❸➌✚➎❁➄✹➋✥➑✿➉❛➈✥➆★➒✇➓ ➏➍➃❸➔✲➈✹➄❛➈✍➉❛➚♠➃❸➊✍➈✹➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈✍➌❤➏❛➄✍❒✦➛❤❮✚➝❤➜➍➝❤↔❩↕❦➙r➙➡➞✿➄❛➈✍➉❛➚♠➃❸➊✍➈✂➆❫➌✚➎✆Ï✰➜❏➙➡➞❤➢❷➤❫↕❦➙r➞ ➄❛➈❄➉❛➚♠➃❸➊✍➈♠➻✿➯➍➌✲➲❼➈❄➊✍➏➍➃❸➋♠➌✠✟Ò➪✖➈✖➅✚➉❛➈✍➄❛➈✍➌❤➏✿➆❳➏➍➃❸➔✵➈➡➓❏➄❛➏r➆❫➔✵➅✲➶❩➆✥➄❛➈✍➎✵➄❛➊r➂❁➈✍➎✚➐❤➒✇➓ ➃❸➌✚Ð⑧➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➪❳➂❼➃❸➊r➂Ñ➃❸➄✲➆✪➌❁➋❭➌❤➐❁➌❤➃✇➑❦➋♠➉❛➔✂➒❸➘Ý➪✦➈➡➃❸Ð♠➂❤➏❛➈✍➎❞➚❭➈✍➉❛➄➍➃❸➋♠➌▼➋✥➑ ➏❛➂✚➈☛✡❙➝❤➢③➜✌☞✖❮✚➙❫❮✚➙✥➢③↔❤➛✖➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➎❼➃❸➄❛➊❄➐❁➄❛➄❛➈✍➎✪➃❸➌❅✬✖➈➡➑❦➈❄➉❛➈✍➌❁➊❄➈✠✍✟✏✎❏➹❩➆✥➌❁➎ ➏❛➂✚➈✍➌☎➉❛➈➡➑❦➈❄➉❢➏❛➋✂➆✠➉❛➈✍➊✍➈✍➌❤➏✿➉❛➈✍➄❛➐❤➒❸➏✖➎❁➐✚➈❳➏❛➋✂➵❢➆❫➉❛➈✍✃❤➂✏➆✥➌❁➎✆✑✹➆✍➒✇➒❷➆❫Ð❭➈✍➉✹➭❦➄❛➈✍➈ ✬✖➈➡➑s➈✍➉❛➈✍➌✚➊✍➈✍➄☛✍✒✂✔✓❤➹✦✮✖✕✏✎❦➺✖➪❳➂❼➃❸➊r➂❋➄❛➏r➆❫➏❛➈❄➄✠➏❛➂➇➆❫➏✍➹☛➐✚➌❁➎✚➈✍➉✠➊❄➈✍➉❛➏r➆★➃❸➌✞➊✍➋♠➌✚➎❤➃✇➓ ➏➍➃❸➋❭➌❁➄✍➹ä➏❛➂❼➃❸➄✿➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✽➎❁➈r➒✇➃❸➚♠➈✍➉❛➄✿Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎✵➄❛➈✍➉❛➚♠➃❸➊❄➈✳➃❸➌✵➆✖➌✚➈✍➏➍➪✦➋♠➉❛✃ ➋✥➑➮➆❫➉❛➶❤➃❸➏❛➉r➆❫➉❛➘❞➏❛➋♠➅❩➋✥➒❸➋♠Ð❭➘♠➻▼➼✞➈☎➏❛➂❁➈✍➌❞➏❛➐✚➉❛➌✧➹❢➃❸➌→➲❤➈✍➊✍➏➍➃❸➋♠➌✚➄✘✗❋➆❫➌✚➎✚✙❼➹ ➏❛➋✏➏❛➂❁➈✲➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❞➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄✵➶❩➈❄➄❛➏✹➄❛➐❼➃❸➏❛➈✍➎✞➑❦➋♠➉✹➅✚➉❛➋✥➚♠➃❸➎❼➃❸➌✚ÐÑ➅✚➉❛➈➡➓ ➎❤➃❸➊✍➏❛➈✍➎❞➄❛➈✍➉❛➚♠➃❸➊❄➈♠➻✪➁❳➂✚➈✵➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎❞➄❛➈✍➉❛➚♠➃❸➊✍➈✪➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð❞➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➃❸➌✚➊✍➋❭➉❛➅❩➋♠➉r➆✥➏❛➈✍➄❳➏➍➪✖➋Ò➌❁➋✥➚♠➈r➒❩➃❸➎❁➈★➆❫➄☛✰✚➏❛➂❩➆✥➏✳➋✥➑☛➐✚➄➍➃❸➌❁Ð✜✛✻➯✢✛✦✫✸➄❛➈✍➉❛➚♠➃❸➊✍➈➴➃❸➌➾➆ ➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✠➊❄➋♠➌❤➏❛➈✍➬❤➏✍➹✚➆✥➌❁➎✂➏❛➂➇➆❫➏✿➋❫➑☛➊✍➋❭➉❛➉❛➈➡➒❷➆❫➏➍➃❸➌✚Ð✂➏❛➂❁➈❳➷❤➐✚➈✍➐❁➈r➃❸➌❁Ð✂➎❁➈r➒❷➆★➘ ➋✥➑✰➆✖➅❩➆✥➊➡✃❭➈✍➏✦➆❫➏✻➄❛➐✚➊✍➊✍➈❄➄❛➄➍➃❸➚♠➈✠➌✚➋❤➎✚➈✍➄✻➃❸➌✯➃❸➏❛➄❢➅➇➆❫➏❛➂✂➏❛➋✠➉❛➈✍➎✚➐❁➊❄➈✖➎❁➈r➒❷➆★➘✾✲➍➃❸➏➍➓ ➏❛➈❄➉✍➻❢➼Ñ➈Ò➊✍➋♠➔✂➶❤➃❸➌❁➈✹➏❛➂✚➈✍➄❛➈✹➏➍➪✦➋✹➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð✏➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✲➄➴➃❸➌❋➲❤➈✍➊✍➏➍➃❸➋❭➌ ✣ ➹❩➅✚➉❛➈✍➄❛➈✍➌❤➏➍➃❸➌✚Ð❋➆✯➐✚➌❤➃✇➽❩➈✍➎✞➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✞➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔❅➪❳➂❤➃❸➊➡➂✏➅❁➉❛➋✥➚♠➃❸➎✚➈✍➄ ➶➇➋♠➏❛➂❞Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎→➆❫➌✚➎❞➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❞➄❛➈✍➉❛➚♠➃❸➊❄➈♠➻✤✑➴➃❸➚♠➈✍➌→➏❛➂❁➈☎➊✍➐✚➉❛➉❛➈✍➌❤➏ ➑s➉❛➈✍➌❁à❄➘✆➋✥➑✿➆✥➊✍➏➍➃❸➚♠➃❸➏➍➘☎➃❸➌❋➏❛➂❁➈Ò➎❁➈✍➄➍➃❸Ð❭➌❋➋❫➑❱➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✂➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❋➆✍➒❸Ð♠➋✥➓ ➉➍➃❸➏❛➂✚➔✵➄✍➹➇➪✖➈✹➎❁➋✲➌❁➋❭➏➴➈❄➬❤➅❩➈✍➊✍➏❳➏❛➂➇➆❫➏❳➏❛➂❁➈✂➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➅❁➉❛➈❄➄❛➈✍➌❤➏❛➈✍➎❋➂✚➈✍➉❛➈ ➪✖➃✇➒✇➒✦➶❩➈☎➏❛➂❁➈✹➽➇➌❩➆✍➒❳➪✖➋♠➉❛➎❋➋❭➌▼➏❛➂✚➈✯➔☎➆✥➏❛➏❛➈✍➉☛✰❢➂✚➋✥➪✖➈✍➚♠➈❄➉✍➹✻➪✖➈✂➎✚➋✪➂✚➋♠➅❩➈ ➏❛➂➇➆❫➏✂➏❛➂✚➈✆➃❸➌✚➄➍➃❸Ð♠➂❤➏☎➈✍➔✂➶❩➋❤➎❼➃❸➈❄➎✸➏❛➂✚➈✍➉❛➈➡➃❸➌❞➪✖➃✇➒✇➒✠➶➇➈✆➋✥➑❳➒❷➆❫➄❛➏➍➃❸➌✚Ð▼➚❭➆★➒❸➐✚➈♠➻ ➯❏➌Ñ➅➇➆❫➉❛➏➍➃❸➊❄➐❼➒❷➆✥➉✍➹✳➪✦➈✵➏❛➂❼➃❸➌✚✃✞➏❛➂❩➆✥➏✹➏❛➂❁➈✯➃❸➌✚➄➍➃❸Ð♠➂❤➏✲➐❁➌✚➎❁➈❄➉➍➒❸➘♠➃❸➌❁Ð▼➋❭➐❁➉✹➎✚➈➡➓ ➄➍➃❸Ð❭➌✧➹ä➏❛➂❩➆✥➏✻➃❸➏✧➃❸➄❢➌✚➈✍➊✍➈✍➄❛➄r➆✥➉❛➘✯➏❛➋❳➎❤➃❸➄❛➏➍➃❸➌❁Ð❭➐❼➃❸➄❛➂✏➶❩➈✍➏➍➪✦➈✍➈✍➌✂➏❛➂❁➈✿➏➍➪✖➋❳➶➇➆❫➄➍➃❸➊ ➅✚➉➍➃❸➌✚➊➡➃❸➅❼➒❸➈❄➄✹➋❫➑❳➢❸➥✍➫✖✥❸➝❤↕❦➢③➫❤↔Ò➆❫➌✚➎❋➥✢✦❩➝❤➜r➢③↔❤➛♠➹✥➃❸➄✖➶❩➋♠➏❛➂☎➑❦➐✚➌❁➎➇➆❫➔✵➈❄➌❤➏r➆★➒✿➆✥➌❁➎ ➌✚➋✥➚♠➈➡➒❦➻ ➯➍➌✂➲❤➈✍➊✍➏➍➃❸➋❭➌✆✧✖➪✖➈✿➉❛➈✍➏❛➐❁➉❛➌✹➏❛➋Ò➏❛➂❁➈❢➃❸➄❛➄❛➐✚➈✖➋❫➑➇➏❛➂❁➈✿➄❛➈✍➉❛➚♠➃❸➊✍➈✖➃❸➌❤➏❛➈✍➉➍➑◗➆❫➊✍➈♠➻ ➲♠➃❸➌✚➊✍➈✪➏❛➂✚➈✵➄❛➈✍➉❛➚♠➃❸➊✍➈☎➃❸➌❤➏❛➈✍➉➍➑❏➆✥➊✍➈☎➪Ô➃✇➒✇➒Ô➶❩➈✵➃❸➌❤➚♠➋♠✃❭➈✍➎❞➶❤➘▼➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌❩➄✍➹ ➪✦➈✹➈✍➬❤➅❩➈✍➊✍➏Ò➏❛➂❩➆✥➏Ò➆✯➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✲➄❛➈✍➉❛➚♠➃❸➊✍➈✹➃❸➌❤➏❛➈✍➉➍➑❏➆❫➊❄➈✯➪✖➃✇➒✇➒✿➋♠➐✚➏➍➒✇➃❸➚♠➈➾➆❫➌❤➘ ➅➇➆❫➉❛➏➍➃❸➊❄➐❼➒❷➆✥➉✪➐✚➌❁➎✚➈✍➉➍➒❸➘♠➃❸➌✚ÐÝ➌✚➈✍➏❏➪✖➋♠➉❛✃▼➔✲➈✍➊➡➂➇➆❫➌❤➃❸➄❛➔✪➻❞➁❳➂❤➐✚➄✍➹✖➪✖➈☎➂➇➆★➚❭➈ ➆✥➏❛➏❛➈✍➔✵➅✚➏❛➈✍➎✪➃❸➌✞➋♠➐✚➉➴➅✚➉❛➋♠➅➇➋♠➄r➆★➒✦➏❛➋✵➅✚➉❛➋❤➎❁➐✚➊✍➈☎➆✥➌✪➃❸➌❤➏❛➈✍➉➍➑❏➆✥➊✍➈✯➪❳➂❤➃❸➊➡➂❋➃❸➄ ➱➇➈✍➬♠➃❸➶❤➒❸➈✠➈✍➌✚➋♠➐✚Ð♠➂✂➏❛➋✹➆✥➊✍➊✍➋❭➔✵➔✵➋❤➎➇➆❫➏❛➈❳➆✖➪✖➃❸➎❁➈✦➚♠➆✥➉➍➃❸➈✍➏➍➘✂➋❫➑➇➄❛➐✚➅❁➅❩➋❭➉❛➏➍➃❸➌❁Ð ➔✲➈✍➊➡➂➇➆❫➌❤➃❸➄❛➔✵➄✍➻★☎✠➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌✵➊❄➋♠➌❤➏❛➉❛➋❫➒✚➅➇➋❫➒✇➃❸➊➡➃❸➈✍➄Ô➆❫➉❛➈✿➎❼➃❸➄❛➊✍➐✚➄❛➄❛➈✍➎☎➶✚➉➍➃❸➈➡➱❩➘ ➃❸➌ ➲❼➈❄➊✍➏➍➃❸➋♠➌✩✓❤➹❳➆❫➌✚➎✸➏❛➂✚➈✪➄❛➐✚➅✚➅❩➋♠➉❛➏✏➋❫➑✹➋❭➏❛➂❁➈❄➉✵➄❛➈✍➉❛➚♠➃❸➊✍➈✞➷❤➐❩➆✍➒✇➃❸➏➍➃❸➈✍➄✪➃❸➄ ➊❄➋❫➚❭➈✍➉❛➈✍➎✪➃❸➌✞➲❤➈✍➊✍➏➍➃❸➋❭➌✪✂✔✕❼➻ ➯➍➌❞➋❭➉❛➎❁➈✍➉✲➏❛➋Ñ➶✚➐❼➃✇➒❸➎❑➐✚➅✸➄❛➐❤Ó☎➊➡➃❸➈❄➌❤➏✆➊❄➋♠➌❤➏❛➈✍➬❤➏✵➏❛➋✞➔✵➈✥➆✥➌❼➃❸➌✚Ð❫➑s➐❼➒✇➒❸➘ ➊❄➋♠➔✵➅➇➆❫➉❛➈→➋♠➐✚➉Ñ➪✖➋❭➉❛✃❑➏❛➋❑➅❁➉❛➈❄➚♠➃❸➋♠➐✚➄➍➒❸➘✔➅✚➐❁➶❤➒✇➃❸➄❛➂❁➈❄➎✾➪✖➋♠➉❛✃❩➹✂➪✖➈▼➎✚➈➡➓ ➒❷➆✍➘✏➏❛➂❁➈✂➎❁➈❄➏r➆★➃✇➒❸➈✍➎✞➎❼➃❸➄❛➊❄➐❁➄❛➄➍➃❸➋❭➌Ý➋✥➑✿➉❛➈➡➒❷➆❫➏❛➈❄➎❋➪✖➋♠➉❛✃✏➐❁➌❤➏➍➃✇➒✦➲❤➈✍➊✍➏➍➃❸➋❭➌✪✂✖✂♠➻ ß❳➋✥➪✖➈✍➚♠➈❄➉✍➹❁➪✦➈➴➪✖➃❸➄❛➂✏➏❛➋✯➌✚➋♠➏❛➈Ò➂❁➈✍➉❛➈✹➏❛➂➇➆❫➏✦➋♠➐✚➉✖➪✖➋♠➉❛✃✲➶❩➋♠➉❛➉❛➋✥➪❳➄❳➂❁➈✥➆✍➚♠➓ ➃✇➒❸➘✸➑s➉❛➋♠➔ ➏❛➂✚➈❋➉r➆❫➅❤➃❸➎❼➒❸➘✫Ð♠➉❛➋✥➪✖➃❸➌❁Ð▼➒✇➃❸➏❛➈✍➉r➆✥➏❛➐❁➉❛➈⑧➋♠➌❑➅✚➉❛➋❫➚♠➃❸➎❤➃❸➌✚Ð❑➉❛➈✥➆★➒✇➓ ➏➍➃❸➔✲➈✯➄❛➈✍➉❛➚♠➃❸➊❄➈P➃❸➌❋➅➇➆❫➊r✃♠➈✍➏❳➌✚➈✍➏➍➪✦➋♠➉❛✃❤➄✍➻✿➯➍➌✏➅❩➆✥➉❛➏➍➃❸➊✍➐❼➒❷➆✥➉✍➹✻➏❛➂✚➈✹➪✖➋❭➉❛✃❤➄➴➋✥➑ ➵❢➆❫➉❛➈❄✃❤➂▼➆✥➌❁➎✫✑✹➆✍➒✇➒❷➆❫Ð❭➈✍➉✆➭✢✍✒✂✔✓❤➹✼✮✬✕✭✎s➺➡➹✯✮❭➆❫➊✍➋❭➶❁➄❛➋❭➌❞➆❫➌✚➎✤✛✧➒❸➋❫➘❤➎→➭✢✍✰✂✱✟✲✎s➺➡➹ ➆❫➌✚➎✴✳✰➆❫à✥➆✥➉✍➹☎ß➴➘❤➔➾➆❫➌☛➹☎➆❫➌✚➎✽➵❢➆✥➊➡➃✇➽❩➊r➃Ñ➭✢✍✒✂✟✮❤➹☛✂✭✝✭✎s➺ ➂➇➆✍➚♠➈→➆★➒✇➒➾➊✍➋♠➌❤➓ ➏❛➉➍➃❸➶✚➐❁➏❛➈✍➎❋➏❛➋✲➋♠➐✚➉✠➎✚➈✍➄➍➃❸Ð♠➌☛➻ ✛✧➃❸➌❩➆✍➒✇➒❸➘♠➹✻➃❸➌✞➲❤➈✍➊✍➏➍➃❸➋♠➌✫✂✟✮❤➹☛➪✖➈✹➊✍➋❭➌❁➊➡➒❸➐✚➎✚➈✵➋♠➐✚➉✠➅➇➆❫➅❩➈❄➉✹➪Ô➃❸➏❛➂✞➆✂➉❛➈➡➓ ➚♠➃❸➈✍➪→➋❫➑☛➋❭➐❁➉✿➉❛➈✍➄❛➐❤➒❸➏❛➄✖➆✥➌❁➎☎➆❳➶❁➉➍➃❸➈r➑✻➎❤➃❸➄❛➊✍➐✚➄❛➄➍➃❸➋♠➌✏➋❫➑☛➉❛➈➡➒❷➆✥➏❛➈✍➎☎➈✍➊✍➋❭➌❁➋❭➔✯➃❸➊ ➃❸➄❛➄❛➐✚➈✍➄✍➻❳➁❳➂❁➈✌☎Ò➅❁➅❩➈❄➌❁➎❤➃❸➬❋➊✍➋♠➌❤➏r➆✍➃❸➌❁➄✹➎✚➈✍➏r➆✍➃✇➒❸➄✹➉❛➈➡➒❷➆❫➏➍➃❸➌✚Ð✪➏❛➋✵➏❛➂✚➈✹➄➍➃❸➔✂➐❼➓ ➒❷➆❫➏➍➃❸➋❭➌Ñ➉❛➈❄➄❛➐❼➒❸➏❛➄Ò➏❛➂❩➆✥➏✠➆✥➉❛➈✹➅❁➉❛➈❄➄❛➈✍➌❤➏❛➈✍➎✪➃❸➌✞➲❤➈✍➊✍➏➍➃❸➋♠➌✚➄✵✗★➓ ✣ ➻ ✶ ✷✰❾◗Ù✹✸✻✺❤❾❦❽❛Ü✼✺♠❻ÒÙ✾✽❀✿❁✺❭❿❃❂ ❄✢❅✖Ü✒❆✵✺✠❅❢❾❦❿✖❇✪➀ ✶❉❈ Õ ❹✩❊❋❂ ❿❤❻➡❻❳Ù●✽✌✿❍✺♠❿❃❂ ❄■❅✖Ü✒❆☛✺✏❹❏✸❃✸✻❂ Ü✇➀✥❿❤❽❛Ü✇Ù➇Ø❼❻ ➯➍➌☎➏❛➂✚➈✠➎❤➃❸➄❛➊✍➐✚➄❛➄➍➃❸➋♠➌❋➏❛➂➇➆❫➏✳➑❦➋✥➒✇➒❸➋✥➪➴➄✍➹➇➪✖➈✖➑❦➋❤➊✍➐✚➄✖➋♠➌✏➆✹➅❩➆✥➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✠➊r➒❷➆❫➄❛➄ ➋❫➑✂➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈⑧➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃③➋❭➌❁➄✞➪❳➂❼➃❸➊r➂❑➪✖➈❋➎✚➐❁➶✸Ï❍✥❸➝●❑✔▲✢▼r➝❤➤r➧✆➆✥➅❁➅❤➒✇➃❸➊✥➆✍➓ ➏➍➃❸➋♠➌✚➄✍➻✾➯❏➌❑➆▼➅❼➒❷➆✍➘♠➓❏➶➇➆❫➊r✃✛➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌✧➹✲➏❛➂❁➈✞➄❛➋♠➐✚➉❛➊✍➈▼➏r➆✥✃♠➈✍➄✏➄❛➋♠➔✵➈ ➄➍➃❸Ð♠➌➇➆★➒❦➹➇➅❩➆✥➊➡✃❭➈✍➏➍➃❸à✍➈✍➄✖➃❸➏✍➹✚➆✥➌❁➎☎➏❛➂✚➈✍➌✵➏❛➉r➆✥➌❁➄❛➔✯➃❸➏❛➄✿➃❸➏✖➋✥➚♠➈✍➉✿➏❛➂❁➈❳➌✚➈✍➏➍➪✦➋♠➉❛✃❩➻ ➁❳➂❁➈❳➌✚➈✍➏➍➪✦➋♠➉❛✃✯➃❸➌✚➈✍➚♠➃❸➏r➆❫➶❤➒❸➘✆➃❸➌❤➏❛➉❛➋❤➎✚➐✚➊✍➈✍➄✠➄❛➋❭➔✵➈❳➚❭➆❫➉➍➃❷➆❫➏➍➃❸➋❭➌✪➃❸➌☎➏❛➂❁➈❳➎✚➈➡➓ ➒❷➆✍➘✯➋✥➑✰➈✥➆❫➊r➂✯➎✚➈➡➒✇➃❸➚❭➈✍➉❛➈✍➎✪➅➇➆❫➊r✃♠➈❄➏✍➻✳➁❳➂❤➃❸➄✳➚❭➆✥➉➍➃❷➆❫➏➍➃❸➋♠➌☎➂➇➆❫➄✿➏❛➉r➆❫➎❤➃❸➏➍➃❸➋♠➌➇➆★➒✇➒❷➘ ➶❩➈✍➈❄➌Ý➊★➆★➒✇➒❸➈✍➎✤◆✍➢③↕❦↕❦➙❫➜❛➻✏➁➴➂✚➈☎➉❛➈✍➊✍➈➡➃❸➚❭➈✍➉✵➎✚➈✍➅➇➆❫➊r✃♠➈✍➏➍➃❸à✍➈❄➄✆➏❛➂✚➈✪➎➇➆❫➏r➆❋➆✥➌❁➎ ➏❛➂✚➈✍➌✵➆✥➏❛➏❛➈✍➔✵➅✚➏❛➄❢➏❛➋✖➑❏➆★➃❸➏❛➂❤➑❦➐❤➒✇➒❸➘✵➅❼➒❷➆✍➘✹➶➇➆❫➊r✃✹➏❛➂❁➈✦➄➍➃❸Ð♠➌➇➆★➒❦➻✿➁❳➂❼➃❸➄✻➃❸➄❢➎✚➋♠➌✚➈ ➶❤➘▼➶✚➐❼Þ❙➈✍➉➍➃❸➌❁Ð⑧➏❛➂❁➈✵➃❸➌✚➊✍➋♠➔✯➃❸➌✚Ð❞➎❩➆✥➏r➆✪➏❛➋❋➉❛➈✍➔✵➋✥➚♠➈✏➏❛➂❁➈☎➌✚➈✍➏➍➪✦➋♠➉❛✃Ñ➃❸➌❤➓ ➎✚➐❁➊✍➈❄➎❂✲➍➃❸➏❛➏❛➈❄➉❢➆❫➌✚➎✂➏❛➂❁➈✍➌✹➉❛➈❄➅❼➒❷➆✍➘♠➃❸➌✚Ð✯➏❛➂✚➈✳➄➍➃❸Ð❭➌❩➆✍➒☛➆✥➏✻➄❛➋❭➔✵➈✳➎✚➈✍➄➍➃❸Ð❭➌❩➆✥➏❛➈✍➎ Ï❍✥❸➝●❑✲▲■▼➡➝❤➤r➧✖➅❩➋✥➃❸➌❤➏✍➻❏☎✠➌❤➘☎➎➇➆❫➏r➆✵➏❛➂➇➆❫➏Ò➆❫➉❛➉➍➃❸➚❭➈✍➄✠➶❩➈r➑❦➋♠➉❛➈✹➃❸➏❛➄Ò➆❫➄❛➄❛➋❤➊➡➃❷➆✥➏❛➈✍➎ Ï❍✥❸➝●❑✲▲■▼➡➝❤➤r➧❳➅❩➋❫➃❸➌❤➏✂➊✥➆❫➌❋➶➇➈✵➐❁➄❛➈❄➎Ñ➏❛➋☎➉❛➈✍➊✍➋❭➌❁➄❛➏❛➉❛➐✚➊✍➏✂➏❛➂❁➈✂➄➍➃❸Ð♠➌➇➆★➒✆✰✿➎➇➆❫➏r➆ ➆❫➉❛➉➍➃❸➚♠➃❸➌✚Ð☎➆✍➑❦➏❛➈✍➉✿➏❛➂❁➈✖Ï❍✥❸➝●❑✲▲■▼➡➝❤➤r➧✻➅❩➋✥➃❸➌❤➏✖➃❸➄✳➐✚➄❛➈➡➒❸➈✍➄❛➄✳➃❸➌☎➉❛➈✍➊❄➋♠➌✚➄❛➏❛➉❛➐❁➊✍➏➍➃❸➌✚Ð ➏❛➂✚➈✯➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈✵➄➍➃❸Ð♠➌➇➆★➒❦➻❖✛❁➋♠➉Ò➏❛➂❁➈✹➅✚➐❁➉❛➅➇➋♠➄❛➈✍➄✂➋❫➑✿➏❛➂❤➃❸➄✹➅❩➆✥➅❩➈✍➉✍➹☛➪✖➈✂➆✥➄➍➓ ➄❛➐✚➔✵➈✵➏❛➂➇➆❫➏✲➆★➒✇➒✦➄❛➐❁➊r➂❞➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄☎➂➇➆★➚❭➈☎➄❛➐❼Ó☎➊r➃❸➈✍➌❤➏✯➶✚➐❤Þ☛➈✍➉➍➃❸➌✚ÐÑ➏❛➋ ➄❛➏❛➋♠➉❛➈➮➆★➒✇➒☛➅➇➆❫➊r✃♠➈❄➏❛➄✖➪❳➂❼➃❸➊r➂✏➆❫➉❛➉➍➃❸➚♠➈✹➶➇➈➡➑❦➋♠➉❛➈❳➏❛➂✚➈✠➅❤➒❷➆✍➘♠➓➍➶➇➆❫➊r✃☎➅❩➋❫➃❸➌❤➏☛✰➇➪✖➈ ➉❛➈✍➏❛➐✚➉❛➌❋➏❛➋✯➏❛➂❤➃❸➄✠➅➇➋❫➃❸➌❤➏✠➃❸➌❋➲❤➈✍➊✍➏➍➃❸➋❭➌P✂✭✕❼➻ ➸➴➋♠➏➮➆★➒✇➒✿➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈☎➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❤➌✚➄☎➆❫➉❛➈✹Ï❍✥❸➝●❑✲▲■▼➡➝❤➤r➧Ò➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✚➄ ➭③➑❦➋❭➉✖➈✍➬✚➆❫➔✲➅❼➒❸➈♠➹☛➋❭➌❁➈✹➔✂➃❸Ð♠➂❤➏➴➃❸➔➾➆❫Ð❫➃❸➌✚➈✲➆✯➚♠➃❸➄❛➐❩➆✍➒✇➃❸à✥➆❫➏➍➃❷➋♠➌①➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌ ➪❳➂❼➃❸➊r➂✸➔✲➈✍➉❛➈➡➒❸➘▼➎❤➃❸➄❛➅❼➒❷➆✍➘❭➈✍➎⑤➏❛➂✚➈✵➃❸➔☎➆❫Ð❭➈❋➈✍➌✚➊✍➋❤➎❁➈❄➎Ý➃❸➌❞➈✥➆❫➊r➂✸➅➇➆❫➊r✃♠➈❄➏ ➪❳➂❁➈❄➌❁➈✍➚❭➈✍➉✳➃❸➏❳➆✥➉❛➉➍➃❸➚♠➈✍➎➇➺➡➻❳ß❳➋✥➪✖➈✍➚♠➈❄➉✍➹❤➪✖➈❳➶❩➈➡➒✇➃❸➈✍➚❭➈✯➏❛➂✚➈❳➚❭➆❫➄❛➏✖➔☎➆☛✲❛➋❭➉➍➃❸➏➍➘ ➋❫➑Ò➑❦➐✚➏❛➐❁➉❛➈✏➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✞➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄✍➹❳➃❸➌❁➊➡➒❸➐✚➎❤➃❸➌❁Ð→➔✵➋♠➄❛➏✵➚♠➃❸➎✚➈✍➋❞➆✥➌❁➎ ➆❫➐✚➎❤➃❸➋✂➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✚➄❫➹✚➪✖➃✇➒✇➒✚➽❩➏✰➏❛➂❼➃❸➄❢➅➇➆❫➉r➆✥➎❼➃❸Ð❭➔✪➻★✛❁➐❁➉❛➏❛➂✚➈✍➉❛➔✵➋❭➉❛➈♠➹❤➌✚➋♠➌❤➓ ➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃✫➆❫➅✚➅❼➒✇➃❸➊★➆❫➏➍➃❸➋♠➌➇➄Ñ➊✥➆✥➌✸➄❛➏➍➃✇➒✇➒✲➐❁➄❛➈❋➏❛➂✚➈❋➉❛➈✥➆✍➒✇➓➍➏➍➃❸➔✲➈▼➌❁➈❄➏➍➪✖➋❭➉❛✃ ➄❛➈✍➉❛➚♠➃❸➊✍➈✏➅❁➉❛➋✥➚♠➃❸➎✚➈✍➎✸➶❤➘✞➋♠➐✚➉☎➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈♠➹✦➆★➒❸➏❛➂✚➋♠➐✚Ð♠➂→➏❛➂❼➃❸➄✵➄❛➈❄➉❛➚♠➃❸➊✍➈ ➃❸➄✠➌✚➋♠➏❳➄❛➅❩➈❄➊➡➃✇➽❩➊★➆★➒✇➒❸➘✞➏r➆★➃✇➒❸➋♠➉❛➈❄➎Ñ➏❛➋✵➏❛➂✚➈➡➃❸➉Ò➌❁➈✍➈❄➎❁➄✍➻ ➵❱➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃❋➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈☎➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄✲➂❩➆✍➚♠➈✹➄❛➈❄➚♠➈✍➉r➆✍➒✿➄❛➈✍➉❛➚♠➃❸➊✍➈✂➉❛➈➡➓ ➷❤➐❤➃❸➉❛➈✍➔✵➈✍➌❤➏❛➄❋➏❛➂➇➆❫➏☎➃❸➌❼➑❦➋❭➉❛➔❰➋♠➐✚➉✪➎✚➈✍➄➍➃❸Ð❭➌⑤➅✚➉❛➋♠➅❩➋❭➄r➆★➒❦➻◗✛✧➃❸➉❛➄❛➏✍➹Ò➄➍➃❸➌✚➊✍➈ ➏❛➂✚➈✍➉❛➈ ➃❸➄✏➋❫➑❦➏❛➈❄➌❑➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈Ñ➃❸➌❤➏❛➈❄➉r➆❫➊✍➏➍➃❸➋❭➌✛➶❩➈✍➏❏➪✖➈✍➈✍➌→➏❛➂❁➈✞➏➍➪✖➋⑧➈✍➌❁➎✚➄ ➋❫➑❱➆✥➌✪➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✻➹✰➆✥➄ ➃❸➌ ➆✠➚♠➋❫➃❸➊❄➈✹➊✍➋♠➌❤➚❭➈✍➉❛➄r➆❫➏➍➃❸➋❭➌✧➹➇➏❛➂❁➈➮➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌ ➅❩➈✍➉➍➑s➋♠➉❛➔☎➆✥➌❁➊✍➈✿➃❸➄✻➄❛➈✍➌✚➄➍➃❸➏➍➃❸➚♠➈❳➏❛➋❳➏❛➂❁➈✿➎➇➆❫➏r➆✖➎❁➈r➒✇➃❸➚♠➈✍➉❛➘✂➎❁➈r➒❷➆★➘ ✰❭➃❸➌✯Ð❭➈✍➌❁➈❄➉r➆★➒ ➒❸➋✥➪✖➈✍➉✹➎✚➈➡➒❷➆✍➘✪➃❸➄✹➔✯➐✚➊r➂Ñ➅✚➉❛➈➡➑❦➈✍➉r➆✥➶❼➒❸➈❭➻✵➲❤➈✍➊✍➋❭➌❁➎☛➹☛➃❸➌Ñ➋❭➉❛➎❁➈✍➉✹➏❛➋☎➄❛➈✍➏Ò➏❛➂❁➈ ➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃▼➅❩➋✥➃❸➌❤➏✍➹✳➏❛➂✚➈✏➆❫➅✚➅❼➒✇➃❸➊★➆❫➏➍➃❸➋♠➌❑➌✚➈✍➈✍➎✚➄✂➏❛➋❋➂❩➆✍➚♠➈☎➄❛➋❭➔✵➈✵➃❸➌❼➑❦➋❭➉➍➓ ➔☎➆✥➏➍➃❸➋♠➌✸➭❦➅✚➉❛➈➡➑❦➈✍➉r➆✥➶❼➒❸➘⑧➆❫➌⑧➆❫➶✚➄❛➋❫➒❸➐✚➏❛➈☎➋♠➉✹➄❛➏r➆✥➏➍➃❸➄❛➏➍➃❸➊✥➆★➒❳➶➇➋♠➐✚➌❁➎➇➺✯➆✥➶❩➋♠➐✚➏ ➏❛➂✚➈✹➎❁➈➡➒❷➆✍➘❤➄❳➏❛➂➇➆❫➏❳➈✥➆❫➊r➂✪➅➇➆❫➊r✃♠➈❄➏➴➪✖➃✇➒✇➒❱➈✍➬❤➅❩➈✍➉➍➃❸➈✍➌✚➊✍➈♠➻✦➁❳➂❼➃❸➉❛➎☛➹❩➄➍➃❸➌✚➊✍➈✲➆★➒✇➒ ➎➇➆❫➏r➆✠➃❸➄➴➶✚➐❤Þ☛➈✍➉❛➈✍➎✏➐❁➌❤➏➍➃✇➒❱➏❛➂❁➈Ò➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃✏➅❩➋❫➃❸➌❤➏✍➹➇➏❛➂❁➈➮➆❫➅✚➅❼➒✇➃❸➊★➆❫➏➍➃❸➋♠➌✞➃❸➄ ➃❸➌✚➎❼➃✇Þ☛➈❄➉❛➈✍➌❤➏✵➆✥➄❳➏❛➋✵➪❳➂❁➈✍➌❋➎➇➆❫➏r➆✯➃❸➄Ò➎❁➈➡➒✇➃❸➚❭➈✍➉❛➈✍➎⑧➆❫➄✖➒❸➋♠➌✚Ð❋➆✥➄✖➃❸➏✹➆❫➉❛➉➍➃❸➚❭➈✍➄ ➶❩➈➡➑s➋♠➉❛➈✂➏❛➂❁➈✂➅❼➒❷➆✍➘♠➓❏➶➇➆❫➊r✃Ñ➅❩➋✥➃❸➌❤➏✢❘❁➻❳➁❳➂❤➃❸➄✹➏❛➐❁➉❛➌✚➄✠➋❭➐❁➏Ò➏❛➋✆➶➇➈✵➆✵➊✍➉❛➐✚➊➡➃❷➆★➒ ➅❩➋✥➃❸➌❤➏✍➹❩➆✥➄✻➃❸➏✿➆★➒✇➒❸➋✥➪❳➄✖➐✚➄❢➏❛➋✠➎✚➈➡➒❷➆✍➘✵➊✍➈✍➉❛➏r➆✍➃❸➌✵➅❩➆✥➊➡✃❭➈✍➏❛➄✿➪➴➂❤➃❸➊r➂☎➆❫➉❛➈✳➃❸➌✲➌❁➋ ➎➇➆❫➌✚Ð♠➈✍➉✂➋❫➑❳➔✂➃❸➄❛➄➍➃❸➌❁Ð❋➏❛➂✚➈➡➃❸➉✂➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃⑧➅❩➋❫➃❸➌❤➏✹➃❸➌Ñ➑❏➆✍➚♠➋❭➉✹➋❫➑✦➅❩➆✥➊➡✃❭➈✍➏❛➄ ➪❳➂❼➃❸➊r➂→➆❫➉❛➈♠➻✤✛❁➋❭➐❁➉❛➏❛➂☛➹✖➏❛➂✚➈✍➄❛➈☎➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃→➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋❤➌❁➄✏➊✥➆❫➌❞➋✥➑❦➏❛➈✍➌ ➏❛➋❫➒❸➈❄➉r➆❫➏❛➈✂➏❛➂❁➈✠➒❸➋❭➄❛➄✠➋❫➑✦➆✂➊✍➈✍➉❛➏r➆★➃❸➌✪➑s➉r➆❫➊✍➏➍➃❸➋❭➌❋➋❫➑✿➅➇➆❫➊r✃♠➈✍➏❛➄❳➪✖➃❸➏❛➂❋➋♠➌❤➒❸➘❋➆ ➔✯➃❸➌❤➃❸➔☎➆★➒Ò➎❤➃❸➄❛➏❛➋♠➉❛➏➍➃❸➋❭➌Ý➃❸➌❞➏❛➂❁➈☎➄➍➃❸Ð❭➌❩➆✍➒❦➻Ñ➁❳➂✚➈✍➉❛➈➡➑❦➋❭➉❛➈♠➹❢➏❛➂✚➈☎➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃ ➅❩➋✥➃❸➌❤➏✵➌❁➈❄➈✍➎❞➌❁➋❭➏✯➶❩➈☎➄❛➋❋➎✚➈➡➒❷➆✍➘♠➈✍➎→➏❛➂❩➆✥➏✵➆✥➶❁➄❛➋✥➒❸➐❁➏❛➈r➒❸➘Ý➈❄➚♠➈✍➉❛➘▼➅➇➆❫➊r✃♠➈❄➏ ➆❫➉❛➉➍➃❸➚❭➈✍➄✠➶➇➈➡➑❦➋♠➉❛➈❄➂❩➆✥➌❁➎☛➻ ✶❉❈✼✶❙❅❯❚✾✺✞❱✖❿❤❽rÛ✚❾❲✺✵Ù●✽✌❳❏✺❨❂ ❿✖❩ ➁❳➂❁➈❳➎✚➈➡➒❷➆✍➘✯➃❸➌☎➏❛➂❁➈❳➌✚➈✍➏➍➪✦➋♠➉❛✃✵➎✚➈✍➉➍➃❸➚♠➈❄➄✳➑❦➉❛➋♠➔✽➄❛➈✍➚❭➈✍➉r➆★➒❙➊✥➆❫➐✚➄❛➈✍➄✍➻✿➁❳➂❁➈✍➉❛➈ ➃❸➄✹➃❸➌Ñ➅✚➉r➆❫➊✍➏➍➃❸➊❄➈❋➆➾➒❷➆✥➉❛Ð♠➈✵➽➇➬❤➈✍➎Ñ➊❄➋♠➔✵➅❩➋❭➌❁➈❄➌❤➏✵➏❛➋✪➏❛➂✚➈✵➎✚➈➡➒❷➆✍➘♠➹❢➊✥➆✥➐❁➄❛➈✍➎ ❘ ð✻ì✟✘ è ✘ è❳ñ✻ì✥ç❛ê❦ç❳ñ☛ç❃✘ ÿ✏✙➡ô✏❬rç✹í❦ì✥ç✹é➡è❦è❦ò✟✚✖ó✥í▲✘ ô✍ÿ ✠ ✚✖ç❛ÿ✍í✆✘ ô❄ÿ❫ç❛õ☎ó✥ê❦ç✔✙☛✘ ô❄ò❫è✆✩ ÷ ✠ í❦ì❭é➡í✻í❦ì✥ç✳ê❦ç❛ë❛ç✔✘ ✙➡ç❛ê✰ì♠érè❢è◗ò✏❭Òë✔✘ ç❛ÿ❄í❢ö❫ò☛❁✚ç❛ê❦è✔✪
by the propagation of the packet at the speed of lig ht, and adapt imperceptibly by adjusting silent periods In fact he delay in transmission at each switch point waiting for uch adaptive approaches have been employed in packet ize he entire packet to arrive before commencing the next stage voice applications since the early 77's([28]) of transmission.(Cut-through ne tworks avoid this delay by and VsT([ -5]) packet voice protocols, which are currently starting transmission before receipt is com plete; most packet used to transmit voice on the Internet, are living examp networks are not cut-through. )5 dded to this fixed delay is a of such adaptive applications. It is important to note that of delay related to the adjust to the delivered de lays ally limits to this adapt jitter, is what must be bounded and minimized if adequat ability; for instance, once the delay reaches a certain level is to be achieved it becomes difficult to carry out interactive conversations ueueing is a fundament al consequence of the statistical s not her useful dist inct ion be network clients is how sharing that occurs in packet net works. One way to redu tolerant they are to brief interruptions in service. This le vel jitter might be to elimina te th stical behavior of the of tolerance is not just a function of the application, but Indeed, one misconception is that real-time so also of the end users involved. For instance, a video confe annot be bursty (variable in their transmission rate), but ence allowing one surgeon to remotely assist anot her during must transmit at a fixed invariant rate to achieve a real-time an operation will not be tolerant of any service interrup tion, whereas a video conference-based family reunion migh ransmission rates and to take advantage of statist ical sh happily tolerate interruptions in service (as long as it wa is a advantage of packe ch reflected in a cheaper se e) is thus to bound and characterize the burstiness rather than We can thus characterize network clients along tw It d aptive or rigi d, and tolerant or intolerant. It is unlikely The idea of statistical sharing implies that there are in that an intolerant network client is adaptive, since the adap deed se veral sources using the bandwidth; one cannot share tive process will likely lead, in the event of rapidly changing alone. Our approach to real-time traffic thus looks at the network conditions, to a brief interruption in service while aggregation of traffic as fundament al; the network must be the play-back point is re-adjusting. Furthermore, a tolerant t hat client t better service tha client that is rig id is merely losing the chance to improve its itched or tdM delay. Such a combination of tolerance and rigidity would network)and (2)are protected from the potent ially negative probably reflect the lack of adaptive hardware and soft ware effects of sharing(most obviously the disri which we believe will rd enough t the resource).g with a mis-behaving source that overloads aused by s fairly ubiquitous. We are thus led to the prediction t hat there will be t wo dominant classes of traffic in the rk: intolerant and rigid clients, and tolerant and adaptive g with Ddl clients. We predict that these two classes will likely request om the net work, Th pplication, in order to set its play-back point knowl- these basic considerations about delay and how clients deal n the delay, plus an with it have produced a taxonomy of net work clients th formation forms the nucleus of the net work's service spec Before turning to the issue of service commitments, let ation in the service interface (to be discussed more full us note that one of the key differences bet ween realtim applications and the traditional datagram applications lies Some realtime applications will use priori delay in the nature of the offered traffic. Dat a traffic is typically bound advertised by the network to set the play-back point sporadic and editable. In contrast, real-time appli d will keep the play-back point fixed regardless of the cations often have some int rinsic packet generation process ct ual delays experienced. These we dub rigid applications. which is long lasting compared to the end-to-end delays of For other applications, the receiver will measure the net work the individ ual packets. This process is a consequence of the delay experienced by arriving packets and then adaptively specifics of the application; for exam ple the coding algorithm ove the play-back point to the minimal delay that still pro- for video, along with the nature of the image, will determine uces a sufficiently low loss rate. We call such application the packet generation process. Furthermore, this generation daptive. Notice that adaptive applications will typically process can often be characterized as conforming to some have an earlier play-back point than rigid applications, and traffic filter or bound (such as the token bucket filter which thus will suffer less performance degradation due to delay we describe in Section 4). For instance, the traffic gener This is because the clients estimate of the post facto bound ated by certain video codecs can often be characterized by al delay will likely be less than the a priori bound some peak rate of packet generation. When a net work has ed b rk. On the ot her hand e some knowledge of the traffic load it will have to carry, it he adapt ation process is not perfect and may nally can allocate its resources in a much more efficient manner higher loss ra The idea of adaptive applications is not relevant to 3 sirv uit switched net works, which do not have jitter due to queueing. Thus most real-time devices today, like voice and video codecs, are not adaptive. ,ack of widespread experi- particular client, it know beforeh and some characteri- n that adaptive applications will be tion of the traffic that will be offered by that the network to re liably meet it difficult to build. However, early experiments suggest that it is actually rather easy video can be made to adapt by Yet another example of an early adaptive packet voice applicati dropping or replaying a frame as necessary, and voice can is described in Reference [5]
➶❤➘✏➏❛➂❁➈✂➅✚➉❛➋♠➅➇➆❫Ð❤➆✥➏➍➃❸➋♠➌❞➋❫➑✿➏❛➂✚➈✂➅❩➆✥➊➡✃❭➈✍➏✹➆✥➏✠➏❛➂✚➈✯➄❛➅❩➈❄➈✍➎Ñ➋✥➑❢➒✇➃❸Ð♠➂❤➏✍➹✿➆✥➌❁➎ ➏❛➂✚➈✪➎✚➈➡➒❷➆✍➘Ñ➃❸➌❞➏❛➉r➆❫➌✚➄❛➔✯➃❸➄❛➄➍➃❸➋❭➌✫➆❫➏✂➈✥➆❫➊r➂❞➄❛➪Ô➃❸➏❛➊r➂▼➅❩➋✥➃❸➌❤➏☎➪❳➆★➃❸➏➍➃❸➌✚Ð ➑s➋♠➉ ➏❛➂✚➈✖➈✍➌❤➏➍➃❸➉❛➈❳➅➇➆❫➊r✃♠➈✍➏✿➏❛➋Ò➆❫➉❛➉➍➃❸➚❭➈➴➶➇➈➡➑❦➋♠➉❛➈❳➊✍➋❭➔✵➔✵➈✍➌✚➊➡➃❸➌✚Ð✹➏❛➂❁➈✦➌❁➈✍➬❤➏✿➄❛➏r➆❫Ð❭➈ ➋✥➑✖➏❛➉r➆✥➌❁➄❛➔✂➃❸➄❛➄➍➃❸➋♠➌☛➻✪➭✁❢❮✚↕▲❛↕✒✦❩➜➍➫❤❮❭➛✔✦P➌✚➈✍➏➍➪✦➋♠➉❛✃❤➄✂➆✍➚♠➋✥➃❸➎Ñ➏❛➂❤➃❸➄✹➎❁➈r➒❷➆★➘✞➶❤➘ ➄❛➏r➆✥➉❛➏➍➃❸➌❁Ð✹➏❛➉r➆✥➌❁➄❛➔✂➃❸➄❛➄➍➃❸➋♠➌☎➶❩➈➡➑❦➋❭➉❛➈✳➉❛➈❄➊✍➈➡➃❸➅✚➏ ➃❸➄❢➊✍➋❭➔✵➅❤➒❸➈✍➏❛➈✒✰❤➔✵➋❭➄❛➏❢➅❩➆✥➊➡✃❭➈✍➏ ➌✚➈✍➏❏➪✖➋♠➉❛✃❤➄✿➆✥➉❛➈✳➌✚➋♠➏❢➊✍➐✚➏➍➓❏➏❛➂❁➉❛➋❭➐❁Ð❭➂✧➻ ➺☛☎Ò➎❁➎✚➈✍➎✂➏❛➋➴➏❛➂❤➃❸➄✻➽➇➬❤➈✍➎✵➎✚➈➡➒❷➆✍➘✠➃❸➄✿➆ ➚❭➆✥➉➍➃❷➆❫➶❤➒❸➈✠➆✥➔✵➋♠➐✚➌❤➏❢➋❫➑➇➎✚➈➡➒❷➆✍➘✯➉❛➈➡➒❷➆✥➏❛➈✍➎✂➏❛➋➴➏❛➂✚➈✖➏➍➃❸➔✵➈✦➏❛➂❩➆✥➏✻➈✥➆✥➊➡➂✂➅❩➆✥➊➡✃❭➈✍➏ ➄❛➅➇➈✍➌❁➎✚➄✠➃❸➌❋➄❛➈✍➉❛➚♠➃❸➊✍➈✲➷❤➐❁➈✍➐✚➈✍➄❳➃❸➌Ñ➏❛➂✚➈✹➄❛➪✖➃❸➏❛➊➡➂✚➈✍➄✍➻Ò➁➴➂❤➃❸➄✠➚❭➆✥➉➍➃❷➆❫➏➍➃❸➋♠➌☛➹✰➋♠➉ ◆❄➢③↕❦↕◗➙✥➜❛➹❤➃❸➄✠➪❳➂❩➆✥➏✠➔✂➐❁➄❛➏Ò➶❩➈✹➶❩➋❭➐❁➌✚➎❁➈❄➎❞➆❫➌✚➎❋➔✯➃❸➌❤➃❸➔✯➃❸à✍➈✍➎Ñ➃✇➑Ô➆❫➎✚➈✍➷❤➐➇➆❫➏❛➈ ➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✂➄❛➈✍➉❛➚♠➃❸➊✍➈✯➃❸➄❳➏❛➋✵➶❩➈✂➆✥➊➡➂❤➃❸➈✍➚❭➈✍➎✧➻ ✂✠➐✚➈✍➐❁➈r➃❸➌❁Ð✯➃❸➄Ô➆➴➑❦➐❁➌✚➎➇➆❫➔✵➈✍➌❤➏r➆✍➒✰➊✍➋♠➌✚➄❛➈✍➷❤➐✚➈✍➌❁➊❄➈✹➋❫➑✰➏❛➂❁➈Ò➄❛➏r➆❫➏➍➃❸➄❛➏➍➃❸➊✥➆✍➒ ➄❛➂➇➆❫➉➍➃❸➌✚Ð✪➏❛➂➇➆❫➏Ò➋❤➊✍➊✍➐✚➉❛➄➴➃❸➌❋➅➇➆❫➊r✃♠➈✍➏Ò➌❁➈❄➏➍➪✖➋❭➉❛✃❤➄✍➻✾✫✠➌✚➈✹➪❳➆✍➘✆➏❛➋☎➉❛➈✍➎✚➐✚➊✍➈ ✲❏➃❸➏❛➏❛➈✍➉✯➔✂➃❸Ð♠➂❤➏✵➶❩➈✏➏❛➋❋➈➡➒✇➃❸➔✯➃❸➌➇➆❫➏❛➈❋➏❛➂✚➈✪➄❛➏r➆✥➏➍➃❸➄❛➏➍➃❸➊✥➆★➒Ò➶❩➈✍➂➇➆✍➚♠➃❸➋♠➉☎➋❫➑Ò➏❛➂✚➈ ➄❛➋❭➐❁➉❛➊❄➈✍➄✍➻✿➯➍➌✚➎❁➈✍➈❄➎✧➹✚➋♠➌✚➈✠➔✂➃❸➄❛➊✍➋♠➌✚➊✍➈✍➅✚➏➍➃❸➋♠➌✪➃❸➄✖➏❛➂➇➆❫➏✿➉❛➈✥➆✍➒✇➓➍➏➍➃❸➔✲➈✠➄❛➋♠➐✚➉❛➊✍➈✍➄ ➊★➆❫➌✚➌❁➋❭➏✯➶❩➈☎➶✚➐❁➉❛➄❛➏❏➘Ý➭❦➚❭➆✥➉➍➃❷➆❫➶❤➒❸➈✆➃❸➌▼➏❛➂✚➈➡➃❸➉✂➏❛➉r➆❫➌✚➄❛➔✯➃❸➄❛➄➍➃❸➋♠➌❞➉r➆❫➏❛➈★➺➡➹❢➶❁➐✚➏ ➔✂➐✚➄❛➏❢➏❛➉r➆❫➌✚➄❛➔✯➃❸➏❳➆❫➏✿➆✖➽❩➬❤➈❄➎✯➃❸➌❤➚❭➆❫➉➍➃❷➆❫➌❤➏✦➉r➆❫➏❛➈✖➏❛➋➮➆❫➊r➂❼➃❸➈✍➚❭➈✹➆❳➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈ ➄❛➈❄➉❛➚♠➃❸➊✍➈♠➻✖➼✞➈✠➉❛➈✲❛➈✍➊✍➏✖➏❛➂❤➃❸➄✖➃❸➎❁➈★➆ ✰☛➆✍➒✇➒❸➋✥➪Ô➃❸➌✚Ð☎➄❛➋♠➐✚➉❛➊✍➈✍➄❳➏❛➋✂➂❩➆✍➚♠➈Ò➶❁➐✚➉❛➄❛➏➍➘ ➏❛➉r➆✥➌❁➄❛➔✂➃❸➄❛➄➍➃❸➋♠➌✞➉r➆❫➏❛➈✍➄❳➆✥➌❁➎☎➏❛➋✂➏r➆❫✃❭➈✹➆❫➎✚➚❭➆❫➌❤➏r➆✥Ð♠➈✹➋❫➑❱➄❛➏r➆❫➏➍➃❸➄❛➏➍➃❸➊✥➆✍➒❢➄❛➂➇➆❫➉➍➓ ➃❸➌✚Ð✵➃❸➄Ò➆P➔☎➆☛✲❛➋❭➉✠➆✥➎❁➚❭➆✥➌❤➏r➆❫Ð♠➈✂➋❫➑❱➅❩➆✥➊➡✃❭➈✍➏❳➌❁➈❄➏➍➪✖➋❭➉❛✃❤➄✍➻ ✫✠➐✚➉Ò➆❫➅✚➅❁➉❛➋ä➆❫➊r➂ ➃❸➄✿➏❛➂❤➐✚➄❢➏❛➋✠➶❩➋❭➐❁➌✚➎☎➆❫➌✚➎✂➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✍➈❳➏❛➂✚➈✖➶❁➐✚➉❛➄❛➏➍➃❸➌✚➈✍➄❛➄✍➹✚➉r➆❫➏❛➂✚➈✍➉✳➏❛➂➇➆❫➌ ➈r➒✇➃❸➔✯➃❸➌❩➆✥➏❛➈✵➃❸➏✍➻ ➁❳➂❁➈✹➃❸➎✚➈✥➆✵➋❫➑✦➄❛➏r➆❫➏➍➃❸➄❛➏➍➃❸➊✥➆✍➒✖➄❛➂➇➆❫➉➍➃❸➌✚Ð✆➃❸➔✵➅❤➒✇➃❸➈✍➄✂➏❛➂❩➆✥➏✠➏❛➂✚➈✍➉❛➈✲➆❫➉❛➈✹➃❸➌❼➓ ➎✚➈✍➈✍➎✏➄❛➈✍➚♠➈❄➉r➆★➒✰➄❛➋♠➐✚➉❛➊✍➈✍➄Ò➐❁➄➍➃❸➌✚Ð✵➏❛➂✚➈✹➶❩➆✥➌❁➎✚➪✖➃❸➎❁➏❛➂✍✰☛➋♠➌✚➈✹➊✥➆❫➌✚➌✚➋♠➏❳➄❛➂❩➆✥➉❛➈ ➆✍➒❸➋♠➌✚➈♠➻ ✫✠➐❁➉➾➆❫➅✚➅❁➉❛➋ä➆❫➊r➂Ý➏❛➋❋➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✪➏❛➉r➆✍Ó☎➊☎➏❛➂❤➐❁➄✂➒❸➋❤➋♠✃❤➄☎➆✥➏✯➏❛➂✚➈ ➆✥Ð♠Ð❭➉❛➈✍Ð❤➆✥➏➍➃❸➋♠➌▼➋✥➑✖➏❛➉r➆✍Ó☎➊✵➆✥➄➴➑❦➐✚➌✚➎❩➆✥➔✵➈✍➌❤➏r➆★➒✆✰❱➏❛➂❁➈✂➌❁➈❄➏➍➪✖➋❭➉❛✃✪➔✂➐❁➄❛➏Ò➶❩➈ ➄❛➂➇➆❫➉❛➈❄➎✆➃❸➌✪➄❛➐✚➊r➂❋➆✹➪❳➆✍➘✯➏❛➂➇➆❫➏❳➊➡➒✇➃❸➈✍➌❤➏❛➄✹➭✢✂★➺✳Ð♠➈✍➏✦➶❩➈✍➏❛➏❛➈✍➉✖➄❛➈❄➉❛➚♠➃❸➊✍➈✯➏❛➂➇➆❫➌ ➃✇➑Ò➏❛➂✚➈✍➉❛➈☎➪✖➈✍➉❛➈☎➌✚➋❋➄❛➂❩➆✥➉➍➃❸➌❁Ð✸➭❏➆✥➄✠➃❸➌→➆✪➊➡➃❸➉❛➊✍➐❤➃❸➏✵➄❛➪✖➃❸➏❛➊➡➂✚➈✍➎❞➋♠➉✂➁Òá✠â ➌✚➈✍➏❏➪✖➋♠➉❛✃✚➺❢➆✥➌❁➎✪➭ ✮❭➺✻➆✥➉❛➈✖➅❁➉❛➋❭➏❛➈✍➊✍➏❛➈✍➎✂➑❦➉❛➋♠➔✛➏❛➂✚➈✖➅❩➋❭➏❛➈✍➌❤➏➍➃❷➆★➒✇➒❸➘☎➌✚➈✍Ð❤➆✥➏➍➃❸➚♠➈ ➈rÞ☛➈✍➊✍➏❛➄✂➋❫➑❳➄❛➂➇➆❫➉➍➃❸➌✚Ð▼➭❦➔✵➋❭➄❛➏✹➋♠➶❤➚♠➃❸➋♠➐✚➄➍➒❸➘❞➏❛➂❁➈✲➎❼➃❸➄❛➉❛➐✚➅❁➏➍➃❸➋❭➌✸➋✥➑❳➄❛➈✍➉❛➚♠➃❸➊✍➈ ➊★➆❫➐✚➄❛➈✍➎✵➶❤➘✹➄❛➂➇➆❫➉➍➃❸➌✚Ð✯➪✖➃❸➏❛➂✲➆➴➔✯➃❸➄➍➓❏➶❩➈✍➂➇➆✍➚♠➃❸➌❁Ð✲➄❛➋♠➐✚➉❛➊✍➈✖➏❛➂➇➆❫➏❢➋✥➚♠➈✍➉➍➒❸➋ä➆❫➎✚➄ ➏❛➂✚➈✹➉❛➈✍➄❛➋❭➐❁➉❛➊✍➈★➺➡➻ ✶❉❈ ✄ ❳❏✺♠❿❃❂ Ü❸Ø✆☎✞✝❳Ü✇❽ ❚✪❳❏✺●❂ ❿✬❩ ☎Ò➌→➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌☛➹✠➃❸➌❞➋♠➉❛➎✚➈✍➉✵➏❛➋✞➄❛➈✍➏✯➃❸➏❛➄✵➅❤➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃❞➅❩➋✥➃❸➌❤➏✆✃❤➌✚➋✥➪✖➒✇➓ ➈❄➎❁Ð❭➈✥➆❫➶❤➒❸➘♠➹❢➌✚➈✍➈✍➎✚➄✹➏❛➋✪✃❤➌✚➋✥➪✛➄❛➋♠➔✵➈✲➶❩➋♠➐✚➌❁➎⑧➋♠➌❋➏❛➂✚➈✵➎❁➈r➒❷➆★➘❭➹✻➅❤➒❸➐❁➄✲➆❫➌ ➈❄➄❛➏➍➃❸➔☎➆❫➏❛➈Ò➋❫➑✰➏❛➂❁➈✖➑❦➉r➆✥➊✍➏➍➃❸➋♠➌☎➋✥➑☛➅❩➆✥➊➡✃❭➈✍➏❛➄✖➔✯➃❸➄❛➄➍➃❸➌✚Ð✵➏❛➂➇➆❫➏✿➶❩➋♠➐✚➌✚➎✧➻✦➁➴➂❤➃❸➄ ➃❸➌❤➑❦➋❭➉❛➔☎➆❫➏➍➃❸➋❭➌✪➑❦➋♠➉❛➔✵➄✿➏❛➂✚➈✠➌❤➐✚➊➡➒❸➈✍➐✚➄➴➋✥➑❢➏❛➂✚➈✠➌✚➈✍➏➍➪✦➋♠➉❛✃ ✳ ➄✳➄❛➈✍➉❛➚♠➃❸➊❄➈✹➄❛➅❩➈✍➊➡➓ ➃✇➽➇➊✥➆✥➏➍➃❸➋♠➌Ñ➃❸➌❋➏❛➂✚➈✯➄❛➈❄➉❛➚♠➃❸➊✍➈✯➃❸➌❤➏❛➈✍➉➍➑❏➆✥➊✍➈☎➭❦➏❛➋✆➶➇➈✹➎❼➃❸➄❛➊✍➐✚➄❛➄❛➈✍➎▼➔✲➋♠➉❛➈✹➑❦➐❼➒✇➒❸➘ ➃❸➌✞➲❤➈✍➊✍➏➍➃❸➋❭➌✞✧♠➺➡➻ ➲❤➋♠➔✵➈✞➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈⑧➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄Ñ➪✖➃✇➒✇➒✂➐✚➄❛➈✞➆❫➌✽➝✪Ï✰➜r➢③➫❤➜➡➢✿➎❁➈r➒❷➆★➘ ➶➇➋♠➐✚➌❁➎✞➆✥➎❁➚❭➈✍➉❛➏➍➃❸➄❛➈✍➎❋➶❤➘☎➏❛➂❁➈✹➌✚➈✍➏❏➪✖➋♠➉❛✃✵➏❛➋✲➄❛➈✍➏✖➏❛➂✚➈✹➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃✏➅❩➋❫➃❸➌❤➏ ➆✥➌❁➎✫➪✖➃✇➒✇➒✵✃❭➈✍➈✍➅❑➏❛➂✚➈Ñ➅❤➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃✫➅❩➋❫➃❸➌❤➏❋➽➇➬❤➈✍➎❑➉❛➈✍Ð❤➆✥➉❛➎❼➒❸➈❄➄❛➄Ñ➋✥➑✵➏❛➂✚➈ ➆✥➊✍➏❛➐➇➆★➒❢➎✚➈➡➒❷➆✍➘❤➄❳➈✍➬❤➅❩➈✍➉➍➃❸➈❄➌❁➊✍➈❄➎✧➻❳➁❳➂✚➈✍➄❛➈✹➪✖➈✠➎✚➐✚➶▼➜r➢✇➛❤➢③➞✠➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌❩➄✍➻ ✛✚➋♠➉❢➋♠➏❛➂✚➈✍➉✿➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄✍➹➇➏❛➂❁➈✦➉❛➈✍➊✍➈➡➃❸➚❭➈✍➉✳➪✖➃✇➒✇➒➇➔✵➈✥➆✥➄❛➐❁➉❛➈✖➏❛➂✚➈✖➌✚➈✍➏➍➪✦➋♠➉❛✃ ➎✚➈➡➒❷➆✍➘▼➈✍➬❤➅➇➈✍➉➍➃❸➈✍➌✚➊✍➈✍➎✸➶❤➘⑧➆❫➉❛➉➍➃❸➚♠➃❸➌✚Ð✞➅❩➆✥➊➡✃❭➈✍➏❛➄✵➆✥➌❁➎❞➏❛➂✚➈✍➌❞➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈➡➒❸➘ ➔✲➋❫➚❭➈✖➏❛➂✚➈✖➅❼➒❷➆✍➘♠➓❏➶➇➆❫➊r✃✵➅❩➋❫➃❸➌❤➏✿➏❛➋✠➏❛➂✚➈✖➔✯➃❸➌❤➃❸➔☎➆✍➒☛➎❁➈r➒❷➆★➘✂➏❛➂➇➆❫➏❢➄❛➏➍➃✇➒✇➒☛➅✚➉❛➋❫➓ ➎✚➐✚➊✍➈✍➄✂➆✯➄❛➐❤Ó☎➊➡➃❸➈✍➌❤➏➍➒❸➘❋➒❸➋✥➪⑤➒❸➋♠➄❛➄✹➉r➆✥➏❛➈♠➻❳➼✞➈✹➊✥➆★➒✇➒✿➄❛➐✚➊➡➂⑧➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄ ➝❤➞❤➝➡Ï✰↕❦➢③➠✥➙✍➻→➸➴➋❭➏➍➃❸➊✍➈✞➏❛➂❩➆✥➏✪➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈❞➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄Ñ➪✖➃✇➒✇➒➮➏➍➘❤➅❤➃❸➊✥➆★➒✇➒❸➘ ➂➇➆✍➚♠➈✲➆❫➌✏➈✥➆❫➉➍➒✇➃❸➈✍➉Ò➅❼➒❷➆✍➘♠➓❏➶➇➆❫➊r✃❋➅❩➋❫➃❸➌❤➏Ò➏❛➂❩➆✥➌✪➉➍➃❸Ð❫➃❸➎⑧➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄✍➹✿➆✥➌❁➎ ➏❛➂❤➐✚➄✂➪Ô➃✇➒✇➒Ô➄❛➐❼Þ☛➈❄➉P➒❸➈✍➄❛➄✂➅❩➈✍➉➍➑s➋♠➉❛➔☎➆✥➌❁➊✍➈✏➎❁➈❄Ð♠➉r➆❫➎➇➆❫➏➍➃❸➋❭➌❞➎❁➐✚➈☎➏❛➋✪➎✚➈➡➒❷➆✍➘♠➻ ➁❳➂❤➃❸➄✖➃❸➄✖➶❩➈✍➊✥➆✥➐❁➄❛➈✹➏❛➂✚➈✠➊r➒✇➃❸➈✍➌❤➏☛✳ ➄Ò➈✍➄❛➏➍➃❸➔☎➆✥➏❛➈✹➋❫➑✰➏❛➂❁➈❳Ï➇➫♠➥✍↕✆✟✍➝❤➤✥↕❦➫✖➶❩➋❭➐❁➌✚➎ ➋❭➌❞➆❫➊❄➏❛➐❩➆✍➒❳➎❁➈r➒❷➆★➘⑧➪Ô➃✇➒✇➒✦➒✇➃❸✃♠➈➡➒❸➘❞➶❩➈✯➒❸➈❄➄❛➄✵➏❛➂❩➆✥➌▼➏❛➂✚➈Ñ➝✯Ï✰➜r➢③➫❤➜r➢✻➶❩➋❭➐❁➌✚➎ ➅✚➉❛➈➡➓❏➊✍➋❭➔✵➅❁➐✚➏❛➈✍➎✫➶❤➘❞➏❛➂❁➈✞➌❁➈❄➏➍➪✖➋❭➉❛✃❩➻ ✫✠➌✸➏❛➂✚➈❋➋♠➏❛➂✚➈✍➉✏➂❩➆✥➌❁➎☛➹✠➄➍➃❸➌✚➊✍➈ ➏❛➂✚➈✲➆❫➎➇➆❫➅✚➏r➆❫➏➍➃❸➋♠➌❞➅✚➉❛➋❤➊✍➈✍➄❛➄❳➃❸➄✹➌❁➋❭➏✹➅❩➈✍➉➍➑❦➈✍➊❄➏✹➆❫➌✚➎❋➔☎➆✍➘❋➋❤➊✍➊✥➆✥➄➍➃❸➋♠➌➇➆★➒✇➒❸➘ ➄❛➈❄➏➴➏❛➂✚➈✹➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃✪➅❩➋✥➃❸➌❤➏✹➏❛➋❤➋✯➈★➆❫➉➍➒❸➘♠➹✰➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈✵➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❤➌✚➄✂➪Ô➃✇➒✇➒ ➒✇➃❸✃❭➈➡➒❸➘✞➈✍➬❤➅❩➈✍➉➍➃❸➈✍➌✚➊✍➈✲➆✯➂❤➃❸Ð♠➂✚➈✍➉✠➒❸➋♠➄❛➄Ò➉r➆❫➏❛➈♠➻ ➁❳➂❁➈✂➃❸➎❁➈✥➆☎➋❫➑Ô➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈✏➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❤➌✚➄✯➃❸➄✹➌✚➋♠➏✹➉❛➈➡➒❸➈✍➚❭➆✥➌❤➏✯➏❛➋☎➊➡➃❸➉➍➓ ➊❄➐❼➃❸➏✞➄❛➪✖➃❸➏❛➊➡➂✚➈✍➎✛➌✚➈✍➏➍➪✦➋♠➉❛✃❤➄✍➹✂➪❳➂❼➃❸➊r➂✛➎✚➋❑➌❁➋❭➏❋➂❩➆✍➚♠➈❋✲➍➃❸➏❛➏❛➈❄➉Ñ➎✚➐❁➈⑧➏❛➋ ➷❤➐✚➈✍➐✚➈➡➃❸➌✚Ð❁➻❳➁❳➂❤➐❁➄✦➔✵➋♠➄❛➏✖➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✹➎✚➈✍➚♠➃❸➊✍➈✍➄❳➏❛➋❤➎➇➆★➘❭➹❼➒✇➃❸✃❭➈✯➚❭➋❫➃❸➊✍➈✂➆✥➌❁➎ ➚♠➃❸➎✚➈✍➋❋➊✍➋❤➎✚➈✍➊✍➄✍➹✰➆✥➉❛➈✯➌✚➋♠➏✂➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈♠➻✵✳✰➆❫➊r✃❋➋❫➑✿➪✖➃❸➎❁➈❄➄❛➅❁➉❛➈✥➆✥➎▼➈✍➬❤➅❩➈❄➉➍➃✇➓ ➈❄➌❁➊✍➈✦➔☎➆✍➘✠➉r➆★➃❸➄❛➈✦➏❛➂❁➈✖➊❄➋♠➌✚➊✍➈✍➉❛➌✂➏❛➂❩➆✥➏❢➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈Ò➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❤➌✚➄✖➪✖➃✇➒✇➒➇➶❩➈ ➎❤➃✇Ó☎➊✍➐❤➒❸➏✵➏❛➋☎➶✚➐❼➃✇➒❸➎☛➻✏ß➴➋✥➪✖➈✍➚❭➈✍➉✍➹✰➈✥➆❫➉➍➒❸➘❋➈✍➬❤➅➇➈✍➉➍➃❸➔✵➈✍➌❤➏❛➄✵➄❛➐✚Ð♠Ð❭➈✍➄❛➏✹➏❛➂➇➆❫➏ ➃❸➏✹➃❸➄✲➆❫➊✍➏❛➐➇➆★➒✇➒❸➘▼➉r➆✥➏❛➂❁➈❄➉✯➈✥➆✥➄❛➘♠➻✡✠➴➃❸➎✚➈✍➋❋➊✥➆✥➌▼➶❩➈☎➔☎➆✥➎❁➈☎➏❛➋❋➆✥➎❩➆✥➅❁➏✂➶❤➘ ➎✚➉❛➋♠➅✚➅❤➃❸➌❁Ð▼➋❭➉✯➉❛➈❄➅❼➒❷➆✍➘♠➃❸➌✚Ð✸➆✆➑s➉r➆❫➔✵➈✏➆❫➄✂➌✚➈✍➊✍➈✍➄❛➄r➆✥➉❛➘♠➹✦➆❫➌✚➎Ý➚❭➋❫➃❸➊✍➈✏➊✥➆❫➌ ➆❫➎➇➆❫➅✚➏✆➃❸➔✲➅❩➈✍➉❛➊✍➈✍➅✚➏➍➃❸➶❤➒③➘✫➶❤➘❑➆✥➎✣✲❛➐✚➄❛➏➍➃❸➌✚Ð❑➄➍➃✇➒❸➈✍➌❤➏❋➅❩➈✍➉➍➃❸➋❤➎✚➄✍➻✔➯❏➌✸➑❏➆❫➊❄➏✍➹ ➄❛➐✚➊➡➂✏➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈✂➆✥➅❁➅✚➉❛➋❤➆✥➊➡➂✚➈✍➄❳➂➇➆★➚❭➈✠➶❩➈❄➈✍➌☎➈✍➔✵➅❤➒❸➋❫➘❭➈✍➎✆➃❸➌☎➅➇➆❫➊r✃♠➈✍➏➍➃❸à❄➈✍➎ ➚♠➋✥➃❸➊✍➈✞➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄❋➄➍➃❸➌✚➊✍➈❋➏❛➂✚➈✪➈✥➆✥➉➍➒❸➘ ✣ ✕ ✳ ➄✆➭✢✍ ✮✖✝✏✎❦➺❀✰✦➏❛➂❁➈☛✠✠➁ ➭✢✍ ✮✏✎❦➺ ➆❫➌✚➎☞✠✯☎✳➁✽➭✢✍✰✂✭✗✭✎s➺✖➅❩➆✥➊➡✃❭➈✍➏✹➚♠➋✥➃❸➊✍➈✵➅✚➉❛➋♠➏❛➋❤➊✍➋✥➒❸➄✍➹❢➪❳➂❼➃❸➊r➂▼➆✥➉❛➈✵➊✍➐✚➉❛➉❛➈✍➌❤➏➍➒❸➘ ➐✚➄❛➈✍➎Ñ➏❛➋☎➏❛➉r➆✥➌❁➄❛➔✯➃❸➏✹➚❭➋❫➃❸➊✍➈✲➋♠➌❋➏❛➂✚➈✯➯❏➌❤➏❛➈✍➉❛➌❁➈❄➏✍➹❢➆✥➉❛➈P➒✇➃❸➚♠➃❸➌✚ÐÑ➈❄➬✚➆❫➔✵➅❤➒❸➈✍➄ ➋❫➑✿➄❛➐✚➊r➂❋➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈➾➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌❩➄✍✌❩➻✿➯❏➏Ô➃❸➄✖➃❸➔✵➅➇➋♠➉❛➏r➆❫➌❤➏✹➏❛➋✲➌❁➋❭➏❛➈✹➏❛➂❩➆✥➏ ➪❳➂❼➃✇➒❸➈✏➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈✪➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋❤➌❁➄☎➊✥➆✥➌▼➆✥➎✣✲❛➐✚➄❛➏✹➏❛➋☎➏❛➂❁➈✲➎❁➈➡➒✇➃❸➚❭➈✍➉❛➈✍➎❞➎✚➈➡➓ ➒❷➆✍➘❤➄✳➋✥➚❭➈✍➉❢➄❛➋♠➔✵➈✦➉r➆❫➌✚Ð♠➈♠➹❤➏❛➂✚➈✍➉❛➈❳➆✥➉❛➈✖➏➍➘❤➅❤➃❸➊✥➆✍➒✇➒❸➘✯➒✇➃❸➔✯➃❸➏❛➄✖➏❛➋Ò➏❛➂❼➃❸➄✦➆❫➎➇➆❫➅✚➏➍➓ ➆❫➶❤➃✇➒✇➃❸➏➍➘ ✰❢➑❦➋❭➉✠➃❸➌❁➄❛➏r➆✥➌❁➊❄➈♠➹❢➋♠➌✚➊✍➈✵➏❛➂✚➈✂➎❁➈➡➒❷➆✍➘✞➉❛➈✥➆❫➊r➂✚➈✍➄✂➆✵➊✍➈✍➉❛➏r➆★➃❸➌✞➒❸➈✍➚♠➈➡➒❦➹ ➃❸➏❳➶❩➈✍➊✍➋❭➔✵➈✍➄Ò➎❼➃✇Ó☎➊❄➐❼➒❸➏✹➏❛➋✵➊★➆❫➉❛➉❛➘☎➋♠➐✚➏➴➃❸➌❤➏❛➈✍➉r➆✥➊✍➏➍➃❸➚♠➈✂➊✍➋❭➌❤➚♠➈✍➉❛➄r➆✥➏➍➃❸➋♠➌✚➄✍➻ ☎✠➌✚➋♠➏❛➂✚➈✍➉✻➐✚➄❛➈➡➑❦➐❤➒✚➎❼➃❸➄❛➏➍➃❸➌✚➊✍➏➍➃❸➋❭➌✆➶➇➈✍➏➍➪✖➈❄➈✍➌✹➌❁➈❄➏➍➪✖➋❭➉❛✃✠➊➡➒✇➃❸➈✍➌❤➏❛➄❢➃❸➄✻➂✚➋✥➪ ➏❛➋❫➒❸➈❄➉r➆❫➌❤➏❳➏❛➂❁➈❄➘☎➆❫➉❛➈Ò➏❛➋✹➶❁➉➍➃❸➈➡➑❢➃❸➌❤➏❛➈❄➉❛➉❛➐❁➅✚➏➍➃❸➋♠➌✚➄➴➃❸➌✏➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻✿➁❳➂❼➃❸➄✖➒❸➈✍➚❭➈➡➒ ➋❫➑➮➏❛➋❫➒❸➈✍➉r➆✥➌❁➊❄➈ ➃❸➄☎➌✚➋♠➏✗✲❛➐✚➄❛➏✪➆✪➑❦➐✚➌❁➊❄➏➍➃❸➋♠➌❑➋❫➑➮➏❛➂❁➈✞➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✻➹✹➶✚➐✚➏ ➆★➒❸➄❛➋☎➋✥➑❢➏❛➂✚➈✹➈✍➌❁➎❋➐✚➄❛➈✍➉❛➄✖➃❸➌❤➚♠➋✥➒❸➚♠➈✍➎☛➻ ✛❁➋❭➉Ô➃❸➌✚➄❛➏r➆❫➌✚➊✍➈♠➹✰➆✂➚♠➃❸➎❁➈✍➋☎➊✍➋❭➌❼➑s➈✍➉➍➓ ➈✍➌✚➊✍➈✹➆✍➒✇➒❸➋✥➪Ô➃❸➌✚Ð☎➋♠➌✚➈✹➄❛➐❁➉❛Ð❭➈✍➋♠➌☎➏❛➋✂➉❛➈✍➔✵➋❭➏❛➈➡➒❸➘✏➆❫➄❛➄➍➃❸➄❛➏Ò➆❫➌✚➋♠➏❛➂✚➈✍➉✖➎✚➐❁➉➍➃❸➌✚Ð ➆❫➌→➋♠➅❩➈❄➉r➆❫➏➍➃❸➋♠➌✫➪✖➃✇➒✇➒✹➌✚➋♠➏✏➶❩➈❋➏❛➋❫➒❸➈✍➉r➆✥➌❤➏✪➋✥➑✂➆❫➌❤➘→➄❛➈✍➉❛➚♠➃❸➊✍➈❋➃❸➌❤➏❛➈✍➉❛➉❛➐✚➅❼➓ ➏➍➃❸➋♠➌☛➹❤➪❳➂✚➈✍➉❛➈✥➆❫➄✦➆➴➚♠➃❸➎✚➈✍➋✂➊✍➋♠➌❤➑❦➈✍➉❛➈❄➌❁➊✍➈r➓➍➶➇➆❫➄❛➈❄➎➾➑◗➆❫➔✯➃✇➒❸➘✂➉❛➈✍➐✚➌❼➃❸➋❭➌✆➔✂➃❸Ð♠➂❤➏ ➂➇➆❫➅✚➅❼➃✇➒❸➘→➏❛➋❫➒❸➈✍➉r➆✥➏❛➈➾➃❸➌❤➏❛➈❄➉❛➉❛➐❁➅✚➏➍➃❸➋♠➌✚➄✆➃❸➌❞➄❛➈✍➉❛➚♠➃❸➊✍➈✞➭❏➆❫➄✹➒❸➋♠➌✚Ð▼➆✥➄P➃❸➏✂➪❳➆✥➄ ➉❛➈➡➱➇➈✍➊✍➏❛➈✍➎✪➃❸➌✞➆✂➊➡➂✚➈✥➆❫➅➇➈✍➉❳➄❛➈✍➉❛➚♠➃❸➊✍➈✵➉r➆✥➏❛➈✥➺➡➻ ➼Ñ➈❳➊✥➆✥➌✆➏❛➂❤➐✚➄✿➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✍➈✹➌✚➈✍➏➍➪✦➋♠➉❛✃✂➊➡➒✇➃❸➈✍➌❤➏❛➄Ò➆★➒❸➋❭➌❁Ð✂➏➍➪✖➋✂➆✥➬❤➈✍➄✍❒ ➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈✏➋♠➉✹➉➍➃❸Ð✥➃❸➎✧➹Ô➆❫➌✚➎▼➏❛➋❫➒❸➈✍➉r➆✥➌❤➏✯➋❭➉✠➃❸➌❤➏❛➋❫➒❸➈✍➉r➆✥➌❤➏✍➻❋➯➍➏✠➃❸➄✂➐✚➌❼➒✇➃❸✃❭➈➡➒❸➘ ➏❛➂➇➆❫➏✦➆❫➌✹➃❸➌❤➏❛➋❫➒❸➈✍➉r➆✥➌❤➏✖➌❁➈❄➏➍➪✖➋❭➉❛✃✯➊➡➒✇➃❸➈❄➌❤➏✳➃❸➄✿➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈♠➹➇➄➍➃❸➌❁➊❄➈➴➏❛➂✚➈❳➆✥➎❩➆✥➅❼➓ ➏➍➃❸➚♠➈Ò➅❁➉❛➋❤➊❄➈✍➄❛➄❳➪Ô➃✇➒✇➒☛➒✇➃❸✃❭➈➡➒❸➘✆➒❸➈✥➆✥➎✧➹✚➃❸➌☎➏❛➂❁➈Ò➈✍➚♠➈✍➌❤➏✦➋❫➑✰➉r➆❫➅❤➃❸➎❼➒❸➘✏➊➡➂➇➆❫➌✚Ð❫➃❸➌✚Ð ➌✚➈✍➏➍➪✖➋❭➉❛✃❋➊✍➋♠➌✚➎❤➃❸➏➍➃❸➋♠➌✚➄✍➹✖➏❛➋❋➆☎➶❁➉➍➃❸➈r➑✖➃❸➌❤➏❛➈✍➉❛➉❛➐✚➅❁➏➍➃❸➋❭➌Ý➃❸➌▼➄❛➈✍➉❛➚♠➃❸➊❄➈✆➪❳➂❤➃✇➒❸➈ ➏❛➂✚➈✠➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃✪➅❩➋✥➃❸➌❤➏✖➃❸➄➴➉❛➈r➓❛➆✥➎✣✲➳➐❁➄❛➏➍➃❸➌✚Ð❁➻❯✛❁➐❁➉❛➏❛➂✚➈✍➉❛➔✵➋❭➉❛➈♠➹☛➆✹➏❛➋❫➒❸➈❄➉r➆❫➌❤➏ ➊➡➒✇➃❸➈✍➌❤➏❳➏❛➂➇➆❫➏✳➃❸➄✦➉➍➃❸Ð❫➃❸➎✆➃❸➄✦➔✵➈✍➉❛➈➡➒❸➘✵➒❸➋♠➄➍➃❸➌✚Ð✵➏❛➂✚➈✠➊r➂❩➆✥➌❁➊❄➈✠➏❛➋✹➃❸➔✵➅❁➉❛➋✥➚❭➈✠➃❸➏❛➄ ➎✚➈➡➒❷➆✍➘♠➻❋➲❤➐✚➊➡➂→➆✆➊❄➋♠➔✂➶❼➃❸➌➇➆❫➏➍➃❸➋❭➌✸➋❫➑Ô➏❛➋❫➒❸➈✍➉r➆✥➌❁➊✍➈ ➆❫➌✚➎▼➉➍➃❸Ð❫➃❸➎❤➃❸➏➍➘❞➪✖➋❭➐❼➒❸➎ ➅✚➉❛➋♠➶➇➆❫➶❤➒❸➘✆➉❛➈r➱❩➈✍➊❄➏✖➏❛➂❁➈✖➒❷➆✥➊➡✃✵➋✥➑❱➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈✠➂➇➆❫➉❛➎✚➪❳➆✥➉❛➈✹➆❫➌✚➎☎➄❛➋❫➑❦➏❏➪❳➆❫➉❛➈❭➹ ➪❳➂❼➃❸➊r➂☎➪✖➈❳➶❩➈➡➒✇➃❸➈✍➚❭➈✠➪✖➃✇➒✇➒✰➄❛➋❤➋♠➌☎➶❩➈❳➊r➂❁➈✥➆✥➅✏➆❫➌✚➎✆➄❛➏r➆✥➌❁➎➇➆❫➉❛➎☎➈✍➌✚➋♠➐✚Ð♠➂☎➏❛➋ ➶❩➈✍➊❄➋♠➔✵➈✠➑❏➆✍➃❸➉➍➒❸➘✪➐✚➶❼➃❸➷❤➐❤➃❸➏❛➋♠➐✚➄✍➻Ò➼Ñ➈✂➆✥➉❛➈✠➏❛➂❤➐✚➄Ô➒❸➈❄➎❋➏❛➋✯➏❛➂✚➈✹➅❁➉❛➈❄➎❼➃❸➊✍➏➍➃❸➋❭➌ ➏❛➂➇➆❫➏✖➏❛➂✚➈✍➉❛➈✹➪✖➃✇➒✇➒❢➶❩➈Ò➏➍➪✖➋✂➎✚➋♠➔✯➃❸➌➇➆❫➌❤➏Ò➊➡➒❷➆❫➄❛➄❛➈✍➄Ò➋❫➑✰➏❛➉r➆★Ó➾➊✠➃❸➌✪➏❛➂✚➈✠➌✚➈✍➏➍➓ ➪✖➋❭➉❛✃❩❒✻➃❸➌❤➏❛➋❫➒❸➈✍➉r➆✥➌❤➏❳➆❫➌✚➎☎➉➍➃❸Ð❫➃❸➎☎➊➡➒✇➃❸➈✍➌❤➏❛➄✍➹➇➆✥➌❁➎✵➏❛➋✥➒❸➈✍➉r➆❫➌❤➏Ô➆❫➌✚➎☎➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈ ➊➡➒✇➃❸➈✍➌❤➏❛➄✍➻✦➼Ñ➈✹➅✚➉❛➈✍➎❤➃❸➊✍➏✠➏❛➂➇➆❫➏✦➏❛➂❁➈✍➄❛➈✹➏❏➪✖➋✂➊➡➒❷➆❫➄❛➄❛➈✍➄Ò➪✖➃✇➒✇➒✻➒✇➃❸✃♠➈r➒❸➘✪➉❛➈✍➷❤➐✚➈✍➄❛➏ ➚♠➈❄➉❛➘✯➎❤➃✇Þ☛➈✍➉❛➈✍➌❤➏✦➄❛➈✍➉❛➚♠➃❸➊✍➈❳➊✍➋♠➔✲➔✯➃❸➏❛➔✵➈✍➌❤➏❛➄✳➑s➉❛➋♠➔✛➏❛➂✚➈❳➌❁➈✍➏❏➪✖➋♠➉❛✃❩➻❱➁➴➂❤➐✚➄✍➹ ➏❛➂✚➈✍➄❛➈✹➶❩➆✥➄➍➃❸➊✵➊✍➋❭➌❁➄➍➃❸➎✚➈✍➉r➆❫➏➍➃❸➋❭➌❁➄✲➆❫➶➇➋♠➐✚➏✠➎✚➈➡➒❷➆✍➘❋➆❫➌✚➎✪➂✚➋✥➪⑤➊r➒✇➃❸➈✍➌❤➏❛➄✹➎❁➈★➆★➒ ➪✖➃❸➏❛➂Ñ➃❸➏✵➂➇➆✍➚♠➈☎➅❁➉❛➋❤➎✚➐✚➊✍➈✍➎→➆✪➏r➆❫➬❤➋❭➌❁➋❭➔✯➘▼➋✥➑❳➌✚➈✍➏➍➪✖➋❭➉❛✃Ñ➊r➒✇➃❸➈✍➌❤➏❛➄☎➏❛➂❩➆✥➏ Ð♠➐❤➃❸➎✚➈✍➄✹➏❛➂❁➈✹Ð❭➋❤➆✍➒❸➄✠➋❫➑✿➋❭➐❁➉Ò➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈♠➻ ✎✳➈➡➑❦➋❭➉❛➈✵➏❛➐❁➉❛➌❤➃❸➌✚ÐÑ➏❛➋✏➏❛➂❁➈✵➃❸➄❛➄❛➐✚➈✆➋✥➑❳➄❛➈✍➉❛➚♠➃❸➊✍➈☎➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈❄➌❤➏❛➄✍➹ ➒❸➈❄➏ ➐✚➄✆➌✚➋♠➏❛➈❋➏❛➂➇➆❫➏✏➋♠➌✚➈❋➋❫➑➮➏❛➂❁➈✞✃♠➈✍➘❞➎❤➃✇Þ☛➈✍➉❛➈✍➌✚➊✍➈✍➄❋➶➇➈✍➏➍➪✖➈❄➈✍➌✸➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈ ➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄✏➆❫➌✚➎Ñ➏❛➂✚➈✵➏❛➉r➆❫➎❤➃❸➏➍➃❸➋♠➌➇➆★➒Ò➎➇➆❫➏r➆✥Ð♠➉r➆❫➔ ➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄✵➒✇➃❸➈✍➄ ➃❸➌✞➏❛➂❁➈✂➌❩➆✥➏❛➐❁➉❛➈✲➋❫➑✦➏❛➂❁➈✂➋❫Þ☛➈❄➉❛➈✍➎Ñ➏❛➉r➆✍Ó☎➊♠➻✂áÒ➆❫➏r➆✵➏❛➉r➆★Ó➾➊✯➃❸➄✠➏❏➘❤➅❼➃❸➊✥➆✍➒✇➒❸➘ ➄❛➅❩➋❭➉r➆❫➎❤➃❸➊✸➆✥➌❁➎❑➐✚➌✚➅❁➉❛➈✍➎❤➃❸➊✍➏r➆✥➶❼➒❸➈♠➻ ➯❏➌❑➊✍➋♠➌❤➏❛➉r➆✥➄❛➏✍➹✠➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈❞➆❫➅✚➅❤➒✇➃✇➓ ➊✥➆✥➏➍➃❸➋♠➌✚➄✹➋❫➑❦➏❛➈✍➌❋➂➇➆✍➚♠➈✵➄❛➋❭➔✵➈✹➃❸➌❤➏❛➉➍➃❸➌❁➄➍➃❸➊☎➅➇➆❫➊r✃♠➈✍➏✹Ð❭➈✍➌❁➈❄➉r➆❫➏➍➃❸➋♠➌✞➅✚➉❛➋❤➊✍➈✍➄❛➄ ➪❳➂❼➃❸➊r➂❋➃❸➄P➒❸➋❭➌❁Ð☎➒❷➆❫➄❛➏➍➃❸➌✚ÐÑ➊✍➋❭➔✵➅➇➆❫➉❛➈✍➎▼➏❛➋☎➏❛➂✚➈✵➈✍➌✚➎❼➓❏➏❛➋✥➓➍➈❄➌❁➎✞➎❁➈r➒❷➆★➘❤➄✂➋❫➑ ➏❛➂✚➈➴➃❸➌✚➎❼➃❸➚♠➃❸➎❩➐➇➆★➒✦➅➇➆❫➊r✃♠➈✍➏❛➄✍➻✿➁❳➂❼➃❸➄❳➅✚➉❛➋❤➊✍➈✍➄❛➄✖➃❸➄Ò➆P➊❄➋♠➌✚➄❛➈✍➷❤➐❁➈❄➌❁➊✍➈✂➋❫➑❱➏❛➂❁➈ ➄❛➅❩➈✍➊r➃✇➽❩➊✍➄✿➋❫➑✚➏❛➂✚➈✦➆❫➅✚➅❼➒✇➃❸➊✥➆✥➏➍➃❸➋❤➌✛✰✚➑❦➋♠➉☛➈✍➬✚➆✥➔✵➅❤➒❸➈➴➏❛➂✚➈✿➊✍➋❤➎❼➃❸➌✚Ð✹➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➑❦➋❭➉❢➚♠➃❸➎❁➈✍➋✚➹☛➆✍➒❸➋♠➌✚Ð✯➪✖➃❸➏❛➂✵➏❛➂✚➈✖➌➇➆❫➏❛➐✚➉❛➈✠➋❫➑❙➏❛➂❁➈✖➃❸➔☎➆✥Ð♠➈♠➹✚➪✖➃✇➒✇➒☛➎✚➈✍➏❛➈✍➉❛➔✯➃❸➌✚➈ ➏❛➂✚➈➴➅➇➆❫➊r✃♠➈❄➏✖Ð♠➈✍➌✚➈✍➉r➆✥➏➍➃❸➋♠➌✪➅✚➉❛➋❤➊✍➈✍➄❛➄✍➻★✛❁➐✚➉❛➏❛➂✚➈✍➉❛➔✵➋♠➉❛➈❭➹❤➏❛➂❼➃❸➄✦Ð♠➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌ ➅✚➉❛➋❤➊✍➈✍➄❛➄☎➊✥➆❫➌❞➋✥➑❦➏❛➈✍➌❞➶❩➈❋➊r➂❩➆✥➉r➆❫➊✍➏❛➈✍➉➍➃❸à❄➈✍➎❑➆✥➄✵➊✍➋♠➌❤➑❦➋❭➉❛➔✯➃❸➌❁Ð❞➏❛➋✞➄❛➋♠➔✵➈ ➏❛➉r➆★Ó➾➊✠➽❁➒❸➏❛➈✍➉❳➋❭➉➴➶➇➋♠➐✚➌❁➎▼➭❦➄❛➐✚➊➡➂ ➆❫➄❳➏❛➂✚➈✹➏❛➋♠✃♠➈❄➌✪➶✚➐❁➊r✃♠➈✍➏❳➽❁➒❸➏❛➈❄➉➴➪❳➂❤➃❸➊➡➂ ➪✖➈☎➎✚➈✍➄❛➊✍➉➍➃❸➶❩➈❋➃❸➌→➲❼➈✍➊❄➏➍➃❸➋♠➌ ✟ä➺➡➻✫✛❁➋❭➉P➃❸➌✚➄❛➏r➆❫➌✚➊✍➈♠➹❳➏❛➂✚➈✪➏❛➉r➆✍Ó☎➊✪Ð❭➈✍➌✚➈✍➉➍➓ ➆❫➏❛➈❄➎Ñ➶❤➘❋➊✍➈❄➉❛➏r➆★➃❸➌✞➚♠➃❸➎❁➈✍➋❋➊❄➋❤➎❁➈✍➊❄➄✹➊✥➆❫➌✞➋❫➑❦➏❛➈❄➌Ñ➶❩➈✲➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✍➈✍➎⑧➶❤➘ ➄❛➋♠➔✲➈✵➅❩➈✥➆✥✃❋➉r➆❫➏❛➈✂➋❫➑❳➅➇➆❫➊r✃♠➈✍➏✹Ð❭➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌✧➻✲➼✸➂✚➈✍➌▼➆☎➌❁➈❄➏➍➪✖➋❭➉❛✃❋➂❩➆✥➄ ➄❛➋♠➔✲➈✆✃❤➌✚➋✥➪✖➒❸➈✍➎❁Ð❭➈✪➋✥➑✠➏❛➂✚➈☎➏❛➉r➆★Ó☎➊✵➒❸➋ä➆❫➎▼➃❸➏✯➪✖➃✇➒✇➒❳➂➇➆✍➚♠➈✏➏❛➋✪➊✥➆✥➉❛➉❛➘♠➹❢➃❸➏ ➊✥➆✥➌❋➆★➒✇➒❸➋❤➊★➆❫➏❛➈✯➃❸➏❛➄Ò➉❛➈✍➄❛➋♠➐✚➉❛➊✍➈✍➄❳➃❸➌✞➆✯➔✯➐✚➊r➂✪➔✵➋❭➉❛➈✹➈➡Ó☎➊➡➃❸➈✍➌❤➏Ò➔☎➆❫➌✚➌✚➈✍➉✍➻ ✄✑✏❉✺❤❾✓✒♠Ü✇➀✔✺✆❊❢Ù❉❆✜❆✯Ü✇❽❆☛✺❤Ø❤❽❛❻ ✳➒❸➈✥➆❫➉➍➒❸➘❭➹ ➑s➋♠➉✂➆✪➌✚➈✍➏➍➪✖➋❭➉❛✃Ñ➏❛➋✕✔✯➝✥➧❤➙✂➆✆➄❛➈❄➉❛➚♠➃❸➊✍➈✪➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈❄➌❤➏✵➏❛➋❋➆ ➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✹➊➡➒✇➃❸➈❄➌❤➏✍➹❩➃❸➏➴➔✂➐✚➄❛➏➴✃❤➌✚➋✥➪⑤➶➇➈➡➑❦➋♠➉❛➈❄➂❩➆✥➌❁➎❋➄❛➋❭➔✵➈✹➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃✇➓ à✥➆✥➏➍➃❸➋♠➌✞➋❫➑✦➏❛➂❁➈✵➏❛➉r➆✍Ó☎➊✂➏❛➂❩➆✥➏✹➪Ô➃✇➒✇➒✿➶➇➈✵➋❫Þ☛➈❄➉❛➈✍➎Ñ➶❤➘❋➏❛➂➇➆❫➏Ò➊➡➒✇➃❸➈✍➌❤➏✍➻✜✛❁➋❭➉ ➏❛➂✚➈❋➌❁➈✍➏❏➪✖➋♠➉❛✃→➏❛➋▼➉❛➈➡➒✇➃❷➆✥➶❼➒❸➘✖✔✯➙r➙✥↕Ô➃❸➏❛➄☎➄❛➈✍➉❛➚♠➃❸➊❄➈▼➊✍➋♠➔✲➔✯➃❸➏❛➔✵➈✍➌❤➏✍➹Ò➏❛➂❁➈ ✌✁✗ç❛í➇é➡ÿ✥ô✍í❦ì✥ç❛ê❁ç✔❊✥é✞✚✖ó✟✩ ç✻ô★ érÿ✳çré➡ê✆✩ ÷✖é➡õ❭é➡ó✥í✆✘ ✙rç✻ó❭é➡ë❬➡ç❛í ✙rô☛✘ ë❛ç❱é➡ó✥ó✣✩ ✘ ëré➡í✆✘ ô✍ÿ ✘ è❢õ✥ç❛è❦ë❛ê✆✘ ö❭ç❛õ✗✘ ÿ✹æ✻ç★ç❛ê❦ç❛ÿ✥ë❛ç✙✘ ☎✍✚✎✪
client must meet its traffic i.e…its rvice will be conform to the characterization it has passed to the net. We haye thus defined two sorts of real time traffic, which work). Thus, the service commit ment made to a particular differ in terms of the service commit ment they receive. There client is predicated on the traffic commit ment of that client is a third class of traffic that we call datagram traffic, to cated on(besides the obvious requirement that the net work except to promise not to delay or drop packets unnecessar- hardware function properly )C ily( this is somet im lled best eg ort service) One kind of service commitment, which we will call guar We now have the first component of anteed service, de pends on no ot her assumptions. That is the mit ment. The challe if the network hardware is functioning and the client is col is to schedule the packet depart ures at each switch so that forming to its traffic characterization, then the these commitments are met. For the sake of clarity, we first mit ment will be met. not ice that this level of commitment onsider, in Section r how to sched ule guaranteed traffic does not require that any ot her net work clients conform to in a net work carrying only guaranteed traffic. In Sections their traffic commitments. Ruaranteed service is appropri and we then consider how to schedule predicted traffic ate for intolerant and rigid clients, since they need absolute network carrying only predicted traffic. After we ha assurances about the service they receive. assembled the necessary components of our scheduling algo- However guaranteed service essarily appropri rithm we then, in Section Apresent our unified scheduling ate for tolerant and adaptive clients. Adaptive clients, b algorithm which simult aneously handles all three levels of adjusting their play-back point to reflect the delays the service commit ment packets are currently receiving, are gambling that the As we present these scheduling schemes, we also lay the ork service in the near fut ar to that del ground work for the other key pieces of the architecture, the red in the recent past. Any violation of that assumption in specifics of the service interface(which must relate closely the direction of increased delays will result in a brief degra- to the details of the service commit ment) and the method new soU missing the play-back point. The client will then readjust the play-back point upward to reflect the change in service but there will necessarily be some momentary disruption w Sckldlu. ing A: goritkM for Guarantlld Traffic In this sect ion we first describe a traffic filter and then a vice commit ments(based on the bounds on scheduling algorithm that together provide guaranteed ser- bec noring those a priori bounds on delay and adapting to the As discussed briefly in a network client must current delivered servi charact Thus, as long as the applicat ion is gambling that the cent past is a guide to the near future, one might as well at realizes the service commitment We use a particular define a class of service commit ment that makes the same form of traffic characterization called a token bucket filter gamble. Our second kind of service commit ment is called a token bucket filter is characterized by t ameters, a dicted service This le yel of comr rate V and a depth b. One can think of the token bucket ponents. First, as stated above, the net work commits that as filling up with tokens cont inuously at a rate v with b if the past guide to the future, then the net being its maximal depth. Bvery time a packet is generated This p tokens are removed from the bucket, where p is the size the fact that the net work can take into measurements of the traffic load in estimating what kind of filter(V b)if there are always enough tokens in the bucket rvice it can deliver reliably. This arked contras whenever a packet is generated the worst-case analysis that underlies the More precisely, vice commitment. Second, the network attempts to delive ti and pi denoting the generation time and size, respectively ice that will allow the adaptive algorithms to minimize of the Yth packet. We say that this traffic source conform their play-back points. (This is the same as saying that the to a token bucket filter(V,b)of rate V and depth b if th service wil tempt to minimize the post facto delay bound. sequence n: defined by no =b and n:= MIN[,Ni-I Obviously, when the overall network conditions change, the quality of service must also change; the intent of the second Y The quantiti go beys the con减rmth;am nneg at f tokens residing in the bucket after the yth packet leaves network schedules packets For a given traffic generation process, we can define the non- so that the current pos ell under the lo lelay bounds(which are typi- increasing function b(y as the minimal value such that the priori bounds that are part process conforms to a(V,b( filter of the service commit ars, several time stam d algorith Notice that predicted service has built into it very stron algorithms take as input mplicit assumptions about the behavior of other net work clients by assuming that the network conditions will remail oportionment of the link as a relatively unchanged, but involves very few eQplicit assump (where a labels the flows); the result ing dela on the bucket sizes b(V ions about these ot her net work chents: ie, their current One of the first such time-stamp algorith behavior need not be explicit ly characterized in any precise iweueing algorithm introduced in Refere anner.Thus, for predicted service, the network takes steps gorithm was targeted at the traditional data o deliver consistent performance to the client; it avoids the tecture, and so involved no preallocation of resources(and hard problem, which must be faced with guaranteed service, thus had each v= here u denotes the link of trying to compute a priori what that level of delivered
➊r➒✇➃❸➈✍➌❤➏❳➔✯➐✚➄❛➏✳➔✲➈✍➈✍➏✻➃❸➏❛➄✖➏❛➉r➆✍Ó☎➊❳➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏✠➭③➃❦➻ ➈♠➻❸➹♠➃❸➏❛➄✳➏❛➉r➆✍Ó☎➊❳➔✯➐✚➄❛➏ ➊❄➋♠➌❤➑❦➋♠➉❛➔✒➏❛➋▼➏❛➂✚➈❋➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌⑤➃❸➏☎➂➇➆❫➄✪➅➇➆❫➄❛➄❛➈❄➎❑➏❛➋▼➏❛➂❁➈❋➌✚➈✍➏➍➓ ➪✦➋♠➉❛✃✚➺➡➻✖➁❳➂❤➐✚➄✍➹☛➏❛➂❁➈✹➄❛➈❄➉❛➚♠➃❸➊✍➈✵➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈✍➌❤➏Ò➔☎➆❫➎✚➈✂➏❛➋✪➆✂➅❩➆✥➉❛➏➍➃❸➊✍➐❼➒❷➆✥➉ ➊r➒✇➃❸➈✍➌❤➏✖➃❸➄✖➅❁➉❛➈❄➎❼➃❸➊✥➆✥➏❛➈✍➎✪➋❭➌☎➏❛➂❁➈❳➏❛➉r➆★Ó➾➊➴➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏❳➋❫➑✰➏❛➂❩➆✥➏✳➊r➒✇➃❸➈✍➌❤➏✍➻ ➁❳➂✚➈✹➷❤➐❁➈❄➄❛➏➍➃❸➋♠➌✪➃❸➄✍➹❩➪❳➂➇➆❫➏✖➈r➒❸➄❛➈✠➃❸➄✠➏❛➂✚➈✠➄❛➈✍➉❛➚♠➃❸➊✍➈✂➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈✍➌❤➏❳➅✚➉❛➈✍➎❤➃✇➓ ➊★➆❫➏❛➈✍➎✏➋♠➌❋➭❦➶❩➈✍➄➍➃❸➎✚➈✍➄✠➏❛➂✚➈✠➋❭➶❤➚♠➃❸➋♠➐✚➄✠➉❛➈✍➷❤➐❤➃❸➉❛➈✍➔✵➈❄➌❤➏✠➏❛➂➇➆❫➏✖➏❛➂✚➈✠➌✚➈✍➏➍➪✦➋♠➉❛✃ ➂➇➆❫➉❛➎✚➪❳➆✥➉❛➈P➑❦➐✚➌✚➊✍➏➍➃❸➋♠➌❋➅✚➉❛➋♠➅❩➈❄➉➍➒❸➘✚➺✁ ✫✠➌✚➈✖✃♠➃❸➌❁➎☎➋✥➑❩➄❛➈✍➉❛➚♠➃❸➊❄➈➴➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏✍➹✚➪➴➂❤➃❸➊r➂✯➪✖➈✦➪Ô➃✇➒✇➒➇➊✥➆✍➒✇➒✰➛❤❮❁➝❤➜✢▲ ➝❤↔➇↕◗➙r➙r➞P➄❛➈❄➉❛➚♠➃❸➊✍➈♠➹❢➎✚➈✍➅❩➈✍➌✚➎✚➄✵➋♠➌✞➌❁➋❋➋❭➏❛➂❁➈❄➉✂➆❫➄❛➄❛➐✚➔✵➅❁➏➍➃❸➋❭➌❁➄✍➻ ➁➴➂➇➆❫➏✠➃❸➄✍➹ ➃✇➑❱➏❛➂❁➈Ò➌❁➈❄➏➍➪✖➋❭➉❛✃✵➂❩➆✥➉❛➎❁➪Ô➆❫➉❛➈➴➃❸➄✖➑❦➐✚➌✚➊✍➏➍➃❸➋♠➌❤➃❸➌❁Ð✏➆✥➌❁➎✏➏❛➂❁➈❳➊➡➒✇➃❸➈✍➌❤➏❳➃❸➄✖➊✍➋❭➌❼➓ ➑s➋♠➉❛➔✯➃❸➌✚Ð✵➏❛➋✹➃❸➏❛➄✖➏❛➉r➆★Ó☎➊Ò➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌☛➹✻➏❛➂✚➈✍➌☎➏❛➂✚➈✠➄❛➈✍➉❛➚♠➃❸➊✍➈✹➊❄➋♠➔✯➓ ➔✂➃❸➏❛➔✵➈✍➌❤➏✠➪✖➃✇➒✇➒✿➶❩➈✂➔✵➈✍➏✍➻Ô➸➴➋❭➏➍➃❸➊✍➈✯➏❛➂➇➆❫➏Ò➏❛➂❼➃❸➄❳➒❸➈✍➚♠➈r➒✿➋❫➑✿➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈✍➌❤➏ ➎✚➋❤➈✍➄☎↔❩➫❤↕☛➉❛➈✍➷❤➐❤➃❸➉❛➈✵➏❛➂➇➆❫➏✂➆❫➌❤➘✏➋♠➏❛➂✚➈✍➉✹➌✚➈✍➏➍➪✖➋❭➉❛✃✪➊➡➒✇➃❸➈❄➌❤➏❛➄✯➊✍➋❭➌❼➑s➋♠➉❛➔ ➏❛➋ ➏❛➂✚➈➡➃❸➉✂➏❛➉r➆★Ó☎➊✂➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈✍➌❤➏❛➄✍➻ ✑✠➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎▼➄❛➈✍➉❛➚♠➃❸➊✍➈✯➃❸➄✲➆❫➅✚➅✚➉❛➋♠➅✚➉➍➃✇➓ ➆✥➏❛➈✠➑❦➋♠➉✿➃❸➌❤➏❛➋❫➒❸➈✍➉r➆✥➌❤➏✂➆❫➌✚➎✪➉➍➃❸Ð❫➃❸➎❋➊r➒✇➃❸➈✍➌❤➏❛➄✍➹☛➄➍➃❸➌❁➊❄➈✹➏❛➂❁➈✍➘☎➌✚➈✍➈✍➎✞➆✥➶❁➄❛➋✥➒❸➐❁➏❛➈ ➆✥➄❛➄❛➐❁➉r➆✥➌❁➊❄➈✍➄✂➆❫➶❩➋❭➐❁➏❳➏❛➂✚➈✯➄❛➈❄➉❛➚♠➃❸➊✍➈✯➏❛➂✚➈✍➘☎➉❛➈✍➊✍➈r➃❸➚♠➈♠➻ ß➴➋✥➪✖➈❄➚♠➈✍➉✍➹✚Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❋➄❛➈✍➉❛➚♠➃❸➊❄➈P➃❸➄❳➌✚➋♠➏❳➌❁➈❄➊✍➈✍➄❛➄r➆❫➉➍➃✇➒❸➘⑧➆❫➅✚➅✚➉❛➋♠➅✚➉➍➃✇➓ ➆✥➏❛➈✵➑❦➋♠➉✂➏❛➋❫➒❸➈✍➉r➆✥➌❤➏☎➆❫➌✚➎→➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈✏➊➡➒✇➃❸➈✍➌❤➏❛➄✍➻✪☎Ò➎❩➆✥➅❁➏➍➃❸➚❭➈✪➊➡➒✇➃❸➈❄➌❤➏❛➄✍➹✖➶❤➘ ➆✥➎✣✲➳➐❁➄❛➏➍➃❸➌✚Ð✸➏❛➂✚➈➡➃❸➉✪➅❤➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃❑➅❩➋✥➃❸➌❤➏✪➏❛➋▼➉❛➈r➱❩➈✍➊❄➏✪➏❛➂✚➈❋➎❁➈➡➒❷➆✍➘❤➄✏➏❛➂❁➈➡➃❸➉ ➅➇➆❫➊r✃♠➈❄➏❛➄✵➆✥➉❛➈✵➊✍➐✚➉❛➉❛➈✍➌❤➏➍➒❸➘Ñ➉❛➈❄➊✍➈➡➃❸➚♠➃❸➌✚Ð❁➹✦➆❫➉❛➈✵Ðä➆❫➔✂➶❼➒✇➃❸➌✚Ð✞➏❛➂❩➆✥➏✹➏❛➂❁➈✵➌✚➈✍➏➍➓ ➪✦➋♠➉❛✃✵➄❛➈✍➉❛➚♠➃❸➊❄➈P➃❸➌✏➏❛➂❁➈✹➌✚➈✥➆❫➉✿➑❦➐❁➏❛➐✚➉❛➈✹➪✖➃✇➒✇➒❢➶❩➈Ò➄➍➃❸➔✯➃✇➒❷➆❫➉✹➏❛➋✂➏❛➂➇➆❫➏❳➎❁➈r➒✇➃❸➚♠➓ ➈❄➉❛➈✍➎✵➃❸➌✆➏❛➂✚➈✠➉❛➈❄➊✍➈✍➌❤➏✖➅➇➆❫➄❛➏✍➻❋☎✠➌❤➘✂➚♠➃❸➋❫➒❷➆✥➏➍➃❸➋♠➌✞➋❫➑✰➏❛➂➇➆❫➏❳➆✥➄❛➄❛➐❁➔✵➅✚➏➍➃❸➋♠➌☎➃❸➌ ➏❛➂✚➈✹➎❤➃❸➉❛➈✍➊✍➏➍➃❸➋♠➌❋➋✥➑❢➃❸➌❁➊❄➉❛➈✥➆❫➄❛➈✍➎❋➎✚➈➡➒❷➆✍➘❤➄✠➪✖➃✇➒✇➒❱➉❛➈✍➄❛➐❼➒❸➏✖➃❸➌ ➆✯➶❁➉➍➃❸➈r➑✿➎❁➈✍Ð❭➉r➆★➓ ➎➇➆❫➏➍➃❸➋❭➌✛➃❸➌⑤➏❛➂✚➈❞➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌✽✳ ➄⑧➅❩➈✍➉➍➑❦➋❭➉❛➔☎➆❫➌✚➊✍➈→➆❫➄❋➅➇➆❫➊r✃♠➈✍➏❛➄✞➶❩➈✍Ð✥➃❸➌ ➔✂➃❸➄❛➄➍➃❸➌❁Ð▼➏❛➂✚➈☎➅❼➒❷➆✍➘♠➓❏➶➇➆❫➊r✃▼➅❩➋❫➃❸➌❤➏✍➻ ➁➴➂✚➈☎➊➡➒✇➃❸➈✍➌❤➏✵➪✖➃✇➒✇➒❳➏❛➂✚➈✍➌❞➉❛➈✥➆✥➎✣✲❛➐✚➄❛➏ ➏❛➂✚➈✹➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃❋➅❩➋❫➃❸➌❤➏❳➐✚➅❤➪❳➆❫➉❛➎❋➏❛➋✂➉❛➈➡➱➇➈✍➊✍➏❳➏❛➂✚➈✹➊➡➂➇➆❫➌✚Ð♠➈✹➃❸➌✪➄❛➈✍➉❛➚♠➃❸➊❄➈♠➹ ➶✚➐✚➏✪➏❛➂✚➈✍➉❛➈▼➪✖➃✇➒✇➒✂➌❁➈❄➊✍➈✍➄❛➄r➆❫➉➍➃✇➒❸➘✫➶❩➈▼➄❛➋❭➔✵➈✞➔✵➋♠➔✵➈❄➌❤➏r➆❫➉❛➘✸➎❤➃❸➄❛➉❛➐✚➅❁➏➍➃❸➋❭➌ ➃❸➌❞➄❛➈✍➉❛➚♠➃❸➊❄➈♠➻❞➁❳➂❼➃❸➄✵➪✖➃✇➒✇➒Ò➋❤➊✍➊✍➐✚➉✵➈✍➚❭➈✍➌▼➃✇➑✹➏❛➂✚➈☎➌❁➈✍➏❏➪✖➋♠➉❛✃✞➃❸➄✵➔✵➈✍➈✍➏➍➃❸➌✚Ð ➃❸➏❛➄✂➌✚➋♠➔✯➃❸➌➇➆★➒❳➄❛➈❄➉❛➚♠➃❸➊✍➈☎➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏❛➄✪➭❦➶➇➆❫➄❛➈✍➎⑧➋♠➌✞➏❛➂❁➈✵➶➇➋♠➐✚➌❁➎✚➄✵➋♠➌ ➏❛➂✚➈✠➄❛➈❄➉❛➚♠➃❸➊✍➈✥➺➡➹❩➶➇➈✍➊✥➆❫➐✚➄❛➈✂➆❫➌✏➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈✵➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌ ➃❸➄❳➏❏➘❤➅❼➃❸➊✥➆✍➒✇➒❸➘✪➃❸Ð❫➓ ➌✚➋♠➉➍➃❸➌✚Ð❋➏❛➂✚➋♠➄❛➈❋➝✹Ï✰➜➡➢③➫❤➜r➢☛➶❩➋♠➐✚➌❁➎✚➄✹➋♠➌✞➎✚➈➡➒❷➆✍➘▼➆✥➌❁➎⑧➆❫➎➇➆❫➅✚➏➍➃❸➌❁Ð❋➏❛➋☎➏❛➂✚➈ ➊❄➐❁➉❛➉❛➈✍➌❤➏Ò➎❁➈r➒✇➃❸➚♠➈✍➉❛➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈❭➻ ➁❳➂❤➐❁➄✍➹➇➆✥➄Ô➒❸➋❭➌❁Ð➾➆❫➄❳➏❛➂❁➈➮➆❫➅✚➅❼➒✇➃❸➊★➆❫➏➍➃❸➋♠➌❋➃❸➄❳Ð❤➆✥➔✯➶❤➒✇➃❸➌❁Ð✏➏❛➂❩➆✥➏✖➏❛➂✚➈✹➉❛➈➡➓ ➊❄➈✍➌❤➏✵➅➇➆❫➄❛➏✯➃❸➄☎➆✪Ð❭➐❼➃❸➎✚➈❋➏❛➋❋➏❛➂✚➈✪➌✚➈✥➆❫➉✯➑s➐❁➏❛➐✚➉❛➈♠➹✿➋♠➌✚➈✪➔✯➃❸Ð❭➂❤➏☎➆❫➄✂➪✖➈➡➒✇➒ ➎✚➈➡➽➇➌❁➈ ➆✆➊➡➒❷➆✥➄❛➄✵➋❫➑Ô➄❛➈✍➉❛➚♠➃❸➊✍➈✪➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏✯➏❛➂➇➆❫➏✂➔☎➆✥✃♠➈✍➄✂➏❛➂❁➈☎➄r➆✥➔✵➈ Ðä➆❫➔✂➶❼➒❸➈❭➻ ✫✠➐❁➉✂➄❛➈✍➊✍➋❭➌❁➎❞✃♠➃❸➌✚➎✸➋❫➑Ò➄❛➈✍➉❛➚♠➃❸➊❄➈❋➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏✯➃❸➄✵➊★➆★➒✇➒❸➈✍➎ Ï✰➜❏➙➡➞❤➢❷➤❫↕❦➙r➞✯➄❛➈✍➉❛➚♠➃❸➊❄➈♠➻▼➁❳➂❤➃❸➄✯➒❸➈✍➚♠➈r➒❳➋❫➑➮➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏✵➂➇➆❫➄✵➏❏➪✖➋✪➊❄➋♠➔✯➓ ➅➇➋♠➌✚➈✍➌❤➏❛➄✍➻✆✛✧➃❸➉❛➄❛➏✍➹❱➆❫➄Ò➄❛➏r➆❫➏❛➈✍➎⑧➆❫➶❩➋✥➚❭➈♠➹✻➏❛➂✚➈✵➌✚➈✍➏➍➪✦➋♠➉❛✃❋➊✍➋❭➔✵➔✯➃❸➏❛➄✹➏❛➂➇➆❫➏ ➃✇➑➮➏❛➂❁➈☎➅➇➆❫➄❛➏✯➃❸➄➾➆✪Ð♠➐❤➃❸➎❁➈❋➏❛➋❋➏❛➂✚➈✆➑❦➐✚➏❛➐✚➉❛➈♠➹✿➏❛➂❁➈✍➌❞➏❛➂✚➈☎➌❁➈✍➏❏➪✖➋♠➉❛✃⑧➪Ô➃✇➒✇➒ ➔✲➈✍➈✍➏✠➃❸➏❛➄✂➄❛➈✍➉❛➚♠➃❸➊✍➈✪➊r➂➇➆❫➉r➆❫➊❄➏❛➈✍➉➍➃❸à✥➆❫➏➍➃❸➋❭➌✧➻✞➁❳➂❤➃❸➄✯➊✍➋❭➔✵➅❩➋❭➌❁➈✍➌❤➏✲➈✍➔✯➶➇➋❤➎❼➓ ➃❸➈❄➄✵➏❛➂❁➈✵➑❏➆✥➊✍➏✹➏❛➂➇➆❫➏✂➏❛➂❁➈☎➌✚➈✍➏➍➪✖➋❭➉❛✃Ñ➊★➆❫➌❞➏r➆❫✃❭➈➾➃❸➌❤➏❛➋✞➆✥➊✍➊✍➋❭➐❁➌❤➏✵➉❛➈❄➊✍➈✍➌❤➏ ➔✲➈✥➆❫➄❛➐✚➉❛➈✍➔✵➈✍➌❤➏❛➄Ò➋❫➑❱➏❛➂❁➈✹➏❛➉r➆✍Ó☎➊✠➒❸➋❤➆✥➎✪➃❸➌✪➈✍➄❛➏➍➃❸➔☎➆✥➏➍➃❸➌❁Ð✏➪❳➂❩➆✥➏✖✃♠➃❸➌❁➎❋➋✥➑ ➄❛➈❄➉❛➚♠➃❸➊✍➈✪➃❸➏☎➊✥➆❫➌→➎❁➈➡➒✇➃❸➚❭➈✍➉✪➉❛➈r➒✇➃❷➆❫➶❤➒❸➘♠➻❑➁❳➂❼➃❸➄✵➃❸➄✵➃❸➌✸➔➾➆❫➉❛✃♠➈❄➎✸➊✍➋❭➌❤➏❛➉r➆❫➄❛➏ ➏❛➋✹➏❛➂✚➈❳➪✖➋♠➉❛➄❛➏➍➓❏➊✥➆✥➄❛➈Ò➆❫➌➇➆★➒❸➘❤➄➍➃❸➄✠➏❛➂➇➆❫➏✿➐✚➌❁➎✚➈✍➉➍➒✇➃❸➈✍➄Ò➏❛➂❁➈❳Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎☎➄❛➈✍➉➍➓ ➚♠➃❸➊❄➈✵➊✍➋♠➔✲➔✯➃❸➏❛➔✵➈✍➌❤➏✍➻➮➲❼➈❄➊✍➋♠➌✚➎✧➹✰➏❛➂✚➈✯➌✚➈✍➏➍➪✦➋♠➉❛✃❋➆✥➏❛➏❛➈✍➔✵➅✚➏❛➄✠➏❛➋✵➎✚➈➡➒✇➃❸➚❭➈✍➉ ➄❛➈❄➉❛➚♠➃❸➊✍➈✵➏❛➂➇➆❫➏✠➪✖➃✇➒✇➒✦➆✍➒✇➒❸➋❫➪✽➏❛➂✚➈✵➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈☎➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄✂➏❛➋✆➔✂➃❸➌❼➃❸➔✯➃❸à❄➈ ➏❛➂✚➈➡➃❸➉Ò➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃✏➅❩➋❫➃❸➌❤➏❛➄✍➻Ò➭❦➁❳➂❼➃❸➄✖➃❸➄✠➏❛➂✚➈✠➄r➆✥➔✵➈✂➆❫➄❳➄r➆✍➘♠➃❸➌✚Ð✆➏❛➂➇➆❫➏❳➏❛➂✚➈ ➄❛➈❄➉❛➚♠➃❸➊✍➈❳➪Ô➃✇➒✇➒❙➆❫➏❛➏❛➈❄➔✵➅❁➏❢➏❛➋❳➔✯➃❸➌❤➃❸➔✯➃❸à✍➈Ò➏❛➂❁➈✿Ï❩➫♠➥❄↕ ✟✍➝❤➤✥↕❦➫✻➎✚➈➡➒❷➆✍➘✯➶➇➋♠➐✚➌❁➎☛➻✇➺ ✫✠➶❤➚♠➃❸➋❭➐❁➄➍➒❸➘❭➹✻➪❳➂✚➈✍➌✪➏❛➂✚➈✹➋✥➚♠➈✍➉r➆✍➒✇➒❢➌✚➈✍➏➍➪✖➋❭➉❛✃☎➊✍➋♠➌✚➎❼➃❸➏➍➃❸➋❭➌❁➄✂➊r➂❩➆✥➌❁Ð❭➈♠➹❩➏❛➂✚➈ ➷❤➐➇➆★➒✇➃❸➏❏➘✪➋✥➑❢➄❛➈✍➉❛➚♠➃❸➊✍➈✹➔✂➐✚➄❛➏❳➆★➒❸➄❛➋✵➊r➂➇➆❫➌✚Ð♠➈✒✰➇➏❛➂❁➈❳➃❸➌❤➏❛➈✍➌❤➏❳➋❫➑❢➏❛➂✚➈✠➄❛➈❄➊✍➋♠➌✚➎ ➊❄➋♠➔✵➅❩➋❭➌❁➈❄➌❤➏✆➋✥➑Ò➏❛➂❁➈✏➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏✯➃❸➄☎➏❛➂➇➆❫➏✂➪➴➂✚➈✍➌❞➌✚➈✍➏➍➪✖➋❭➉❛✃▼➊✍➋❭➌❼➓ ➎❤➃❸➏➍➃❸➋♠➌✚➄✏➆❫➉❛➈☎➉❛➈➡➒❷➆✥➏➍➃❸➚♠➈➡➒❸➘❞➄❛➏r➆❫➏➍➃❸➊❭➹✳➏❛➂✚➈☎➌❁➈❄➏➍➪✖➋❭➉❛✃Ñ➄❛➊r➂❁➈❄➎❁➐❤➒❸➈✍➄☎➅❩➆✥➊➡✃❭➈✍➏❛➄ ➄❛➋✂➏❛➂➇➆❫➏❳➏❛➂❁➈Ò➊✍➐✚➉❛➉❛➈✍➌❤➏✖Ï❩➫♠➥❄↕ ✟❄➝❤➤❫↕❦➫✖➎✚➈➡➒❷➆✍➘✪➶❩➋❭➐❁➌✚➎✚➄P➭❦➪❳➂❤➃❸➊➡➂ ➆❫➉❛➈Ò➏➍➘❤➅❤➃✇➓ ➊★➆★➒✇➒❸➘☎➪✖➈➡➒✇➒✰➐✚➌❁➎✚➈✍➉✖➏❛➂✚➈➴➒❸➋❭➌❁Ð✥➓➍➏❛➈❄➉❛➔❅➝➴Ï✰➜r➢③➫❤➜r➢❼➶➇➋♠➐✚➌❁➎✚➄❳➏❛➂❩➆✥➏✖➆✥➉❛➈✠➅➇➆❫➉❛➏ ➋✥➑✿➏❛➂❁➈✹➄❛➈❄➉❛➚♠➃❸➊✍➈✯➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏r➺✠➆✥➉❛➈✹➄❛➔☎➆★➒✇➒❦➻ ➸➴➋❭➏➍➃❸➊✍➈❳➏❛➂❩➆✥➏✳➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎☎➄❛➈✍➉❛➚♠➃❸➊✍➈Ò➂❩➆✥➄❢➶❁➐❤➃✇➒❸➏✖➃❸➌❤➏❛➋✠➃❸➏✳➚❭➈✍➉❛➘✂➄❛➏❛➉❛➋♠➌✚Ð ➢✔ÔÏ❍✥❸➢③➤✥➢③↕✿➆✥➄❛➄❛➐❁➔✵➅✚➏➍➃❸➋♠➌✚➄✞➆❫➶➇➋♠➐✚➏✆➏❛➂✚➈✪➶➇➈✍➂❩➆✍➚♠➃❸➋❭➉✪➋❫➑➮➋♠➏❛➂✚➈✍➉✵➌✚➈✍➏➍➪✦➋♠➉❛✃ ➊r➒✇➃❸➈✍➌❤➏❛➄✠➶❤➘☎➆✥➄❛➄❛➐❁➔✂➃❸➌❁Ð✵➏❛➂➇➆❫➏✦➏❛➂❁➈Ò➌❁➈✍➏❏➪✖➋♠➉❛✃✲➊✍➋♠➌✚➎❼➃❸➏➍➃❸➋❭➌❁➄✹➪✖➃✇➒✇➒✰➉❛➈✍➔☎➆✍➃❸➌ ➉❛➈r➒❷➆❫➏➍➃❸➚♠➈r➒❸➘✪➐✚➌❁➊r➂❩➆✥➌❁Ð❭➈✍➎☛➹☛➶❁➐✚➏✳➃❸➌❤➚♠➋✥➒❸➚♠➈✍➄Ò➚♠➈✍➉❛➘✂➑❦➈✍➪✫➙✄✂❄Ï❍✥❸➢❷➤❫➢❷↕❤➆❫➄❛➄❛➐✚➔✵➅❤➓ ➏➍➃❸➋❭➌❁➄✏➆✥➶❩➋♠➐✚➏☎➏❛➂❁➈✍➄❛➈✏➋♠➏❛➂✚➈✍➉✵➌✚➈✍➏➍➪✦➋♠➉❛✃▼➊➡➒✇➃❸➈❄➌❤➏❛➄☛✰➴➃❦➻ ➈♠➻❸➹✳➏❛➂✚➈➡➃❸➉☎➊✍➐✚➉❛➉❛➈✍➌❤➏ ➶➇➈✍➂❩➆✍➚♠➃❸➋❭➉✹➌❁➈✍➈❄➎✪➌✚➋♠➏❳➶❩➈✹➈✍➬❤➅❤➒✇➃❸➊➡➃❸➏➍➒❷➘Ñ➊r➂❩➆✥➉r➆❫➊✍➏❛➈❄➉➍➃❸à✍➈✍➎✪➃❸➌✞➆❫➌❤➘☎➅✚➉❛➈✍➊➡➃❸➄❛➈ ➔➾➆❫➌✚➌❁➈✍➉✍➻✿➁❳➂❤➐✚➄✍➹❫➑❦➋❭➉✻➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎☎➄❛➈✍➉❛➚♠➃❸➊✍➈♠➹ä➏❛➂❁➈✖➌✚➈✍➏❏➪✖➋♠➉❛✃✹➏r➆✥✃♠➈✍➄❢➄❛➏❛➈✍➅✚➄ ➏❛➋✂➎✚➈➡➒✇➃❸➚♠➈✍➉Ò➊✍➋♠➌✚➄➍➃❸➄❛➏❛➈✍➌❤➏Ò➅❩➈✍➉➍➑❦➋❭➉❛➔☎➆❫➌✚➊✍➈✹➏❛➋✹➏❛➂✚➈✹➊➡➒✇➃❸➈✍➌❤➏☛✰❁➃❸➏Ò➆✍➚♠➋✥➃❸➎❁➄❳➏❛➂✚➈ ➂➇➆❫➉❛➎☎➅✚➉❛➋♠➶❤➒❸➈✍➔✪➹✚➪❳➂❼➃❸➊r➂✵➔✂➐❁➄❛➏✿➶❩➈✖➑❏➆✥➊✍➈✍➎✵➪✖➃❸➏❛➂✵Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈✍➎☎➄❛➈✍➉❛➚♠➃❸➊❄➈♠➹ ➋✥➑Ò➏❛➉❛➘♠➃❸➌❁Ð▼➏❛➋❋➊❄➋♠➔✵➅✚➐❁➏❛➈❞➝✵Ï✰➜r➢③➫❤➜➡➢❢➪❳➂➇➆❫➏✂➏❛➂❩➆✥➏✯➒❸➈✍➚♠➈r➒✠➋✥➑Ò➎❁➈➡➒✇➃❸➚❭➈✍➉❛➈✍➎ ➄❛➈✍➉❛➚♠➃❸➊✍➈✂➪✖➃✇➒✇➒✿➶❩➈♠➻ ➼Ñ➈❳➂➇➆★➚❭➈✠➏❛➂❤➐✚➄✳➎✚➈➡➽➇➌❁➈❄➎✆➏❏➪✖➋✹➄❛➋♠➉❛➏❛➄✿➋❫➑❙➉❛➈✥➆★➒☛➏➍➃❸➔✲➈➴➏❛➉r➆✍Ó☎➊♠➹❤➪❳➂❤➃❸➊➡➂ ➎❤➃✇Þ☛➈✍➉✻➃❸➌✂➏❛➈✍➉❛➔✵➄☛➋❫➑➇➏❛➂❁➈✿➄❛➈✍➉❛➚♠➃❸➊✍➈✦➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏✻➏❛➂✚➈✍➘Ò➉❛➈✍➊✍➈➡➃❸➚❭➈♠➻✿➁➴➂✚➈✍➉❛➈ ➃❸➄✏➆Ñ➏❛➂❤➃❸➉❛➎❑➊➡➒❷➆❫➄❛➄☎➋❫➑➮➏❛➉r➆★Ó☎➊❋➏❛➂➇➆❫➏☎➪✖➈✏➊✥➆★➒✇➒➾➞❤➝❤↕❦➝✥➛❼➜❏➝✆✔✔➏❛➉r➆★Ó➾➊♠➹✖➏❛➋ ➪❳➂❼➃❸➊r➂❑➏❛➂✚➈Ñ➌✚➈✍➏➍➪✦➋♠➉❛✃✸➔➾➆❫✃♠➈❄➄✪➌✚➋Ý➄❛➈❄➉❛➚♠➃❸➊✍➈Ñ➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏❛➄✞➆❫➏✏➆★➒✇➒❦➹ ➈✍➬❤➊✍➈❄➅❁➏Ò➏❛➋✆➅✚➉❛➋♠➔✂➃❸➄❛➈✯➌✚➋♠➏Ò➏❛➋✆➎✚➈➡➒❷➆✍➘❋➋♠➉❳➎✚➉❛➋♠➅✞➅❩➆✥➊➡✃❭➈✍➏❛➄✠➐✚➌✚➌❁➈✍➊❄➈✍➄❛➄r➆❫➉➍➓ ➃✇➒❸➘Ñ➭❦➏❛➂❤➃❸➄➴➃❸➄Ò➄❛➋♠➔✵➈❄➏➍➃❸➔✵➈✍➄✠➊★➆★➒✇➒❸➈✍➎✚▼➡➙✍➥❄↕❢➙✆☎✠➫❤➜r↕❩➄❛➈✍➉❛➚♠➃❸➊✍➈✥➺➡➻ ➼Ñ➈❋➌✚➋✥➪❅➂❩➆✍➚❭➈❋➏❛➂❁➈❋➽❩➉❛➄❛➏☎➊✍➋❭➔✵➅❩➋❭➌❁➈✍➌❤➏❋➋✥➑✹➋❭➐❁➉ ➆❫➉❛➊r➂❼➃❸➏❛➈✍➊❄➏❛➐❁➉❛➈♠➹ ➏❛➂✚➈✹➌❩➆✥➏❛➐❁➉❛➈✂➋❫➑✿➏❛➂✚➈✯➄❛➈❄➉❛➚♠➃❸➊✍➈✯➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏✍➻❳➁❳➂❁➈✹➊r➂❩➆✍➒✇➒❸➈✍➌✚Ð♠➈♠➹✰➌❁➋✥➪✹➹ ➃❸➄✹➏❛➋☎➄❛➊➡➂✚➈✍➎✚➐❼➒❸➈☎➏❛➂✚➈✯➅➇➆❫➊r✃♠➈❄➏✹➎❁➈✍➅➇➆❫➉❛➏❛➐✚➉❛➈✍➄✂➆✥➏✠➈✥➆✥➊➡➂✞➄❛➪✖➃❸➏❛➊➡➂✞➄❛➋☎➏❛➂❩➆✥➏ ➏❛➂✚➈✍➄❛➈✹➊✍➋♠➔✲➔✯➃❸➏❛➔✵➈✍➌❤➏❛➄Ò➆❫➉❛➈Ò➔✵➈✍➏✍➻❋✛✚➋♠➉✖➏❛➂✚➈✹➄r➆❫✃❭➈✠➋❫➑❱➊➡➒❷➆❫➉➍➃❸➏❏➘♠➹❩➪✦➈➴➽➇➉❛➄❛➏ ➊✍➋❭➌❁➄➍➃❸➎✚➈✍➉✍➹❳➃❸➌❑➲❤➈✍➊✍➏➍➃❸➋♠➌ ✟✚➹❳➂❁➋✥➪✾➏❛➋▼➄❛➊r➂✚➈✍➎❁➐❤➒❸➈✞Ð♠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎❑➏❛➉r➆★Ó☎➊ ➃❸➌→➆✪➌❁➈❄➏➍➪✖➋❭➉❛✃Ñ➊✥➆✥➉❛➉❛➘♠➃❸➌❁Ð⑧➋♠➌❤➒❸➘▼Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎✸➏❛➉r➆✍Ó☎➊♠➻❋➯❏➌❞➲❤➈✍➊✍➏➍➃❸➋♠➌✚➄ ✗✪➆✥➌❁➎✤✙✆➪✦➈✯➏❛➂✚➈✍➌✞➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉✵➂❁➋✥➪✽➏❛➋☎➄❛➊➡➂✚➈✍➎✚➐❼➒❸➈✏➅❁➉❛➈❄➎❼➃❸➊✍➏❛➈❄➎▼➏❛➉r➆★Ó☎➊ ➃❸➌⑧➆➾➌✚➈✍➏❏➪✖➋♠➉❛✃✏➊✥➆❫➉❛➉❛➘♠➃❸➌✚Ð❋➋♠➌❤➒❸➘❋➅❁➉❛➈❄➎❼➃❸➊✍➏❛➈❄➎▼➏❛➉r➆★Ó☎➊❭➻ ☎❳➑❦➏❛➈✍➉❳➪✖➈✂➂❩➆✍➚♠➈ ➆❫➄❛➄❛➈❄➔✯➶❤➒❸➈✍➎✪➏❛➂✚➈❳➌❁➈❄➊✍➈✍➄❛➄r➆❫➉❛➘☎➊✍➋❭➔✵➅❩➋❭➌❁➈✍➌❤➏❛➄❳➋✥➑☛➋♠➐✚➉✖➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✏➆★➒❸Ð❭➋❫➓ ➉➍➃❸➏❛➂✚➔❅➪✖➈✂➏❛➂❁➈✍➌☛➹❩➃❸➌⑧➲❼➈✍➊❄➏➍➃❸➋♠➌ ✣ ➹☛➅❁➉❛➈✍➄❛➈❄➌❤➏✹➋♠➐✚➉✠➐✚➌❼➃✇➽➇➈✍➎▼➄❛➊r➂❁➈❄➎❁➐❤➒✇➃❸➌❁Ð ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➪❳➂❤➃❸➊➡➂→➄➍➃❸➔✯➐❤➒❸➏r➆❫➌✚➈✍➋❭➐❁➄➍➒❸➘❑➂➇➆❫➌✚➎❼➒❸➈❄➄❋➆★➒✇➒Ò➏❛➂✚➉❛➈✍➈✆➒❸➈❄➚♠➈➡➒❸➄☎➋❫➑ ➄❛➈✍➉❛➚♠➃❸➊✍➈✂➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈✍➌❤➏✍➻ ☎✠➄❳➪✖➈✹➅✚➉❛➈✍➄❛➈✍➌❤➏✠➏❛➂✚➈✍➄❛➈✹➄❛➊r➂❁➈❄➎❁➐❤➒✇➃❸➌❁Ð❋➄❛➊r➂✚➈✍➔✵➈✍➄✍➹☛➪✖➈✂➆✍➒❸➄❛➋✵➒❷➆★➘✏➏❛➂❁➈ Ð♠➉❛➋❭➐❁➌✚➎✚➪✖➋♠➉❛✃☎➑❦➋♠➉✖➏❛➂✚➈✠➋❭➏❛➂❁➈❄➉✖✃♠➈✍➘☎➅❤➃❸➈✍➊✍➈✍➄❳➋✥➑❢➏❛➂✚➈✂➆❫➉❛➊r➂❼➃❸➏❛➈✍➊❄➏❛➐❁➉❛➈♠➹➇➏❛➂❁➈ ➄❛➅❩➈✍➊r➃✇➽❩➊✍➄✲➋❫➑❳➏❛➂✚➈☎➄❛➈✍➉❛➚♠➃❸➊✍➈✆➃❸➌❤➏❛➈❄➉➍➑❏➆❫➊✍➈❋➭❦➪❳➂❤➃❸➊➡➂▼➔✂➐✚➄❛➏✯➉❛➈➡➒❷➆✥➏❛➈☎➊➡➒❸➋♠➄❛➈➡➒❸➘ ➏❛➋✪➏❛➂✚➈☎➎❁➈❄➏r➆★➃✇➒❸➄✵➋✥➑❳➏❛➂❁➈✲➄❛➈✍➉❛➚♠➃❸➊✍➈✪➊❄➋♠➔✵➔✯➃❸➏❛➔✲➈✍➌❤➏r➺✯➆✥➌❁➎⑧➏❛➂❁➈✵➔✲➈✍➏❛➂❁➋❤➎ ➏❛➋✵➊✍➋❭➌❤➏❛➉❛➋❫➒❱➏❛➂❁➈✂➆✥➎❁➔✯➃❸➄❛➄➍➃❸➋❭➌▼➋❫➑❱➌❁➈❄➪⑤➄❛➋❭➐❁➉❛➊✍➈❄➄✍➻ ✝ ✏❩➀✬❚✾✺❭Ú❩Û❃❂ Ü❸Ø✆☎✪❹★❂ ☎❁Ù✚❾❦Ü✇❽ ❚✻❆✯❻❏✽❏Ù✚❾✟✞✖Û❤❿❤❾❦❿✚Ø❼❽✢✺✬✺♠Ú✘❅❢❾◗❿✖❇✏➀ ➯➍➌❞➏❛➂❤➃❸➄☎➄❛➈✍➊✍➏➍➃❸➋♠➌→➪✖➈✆➽➇➉❛➄❛➏✵➎✚➈✍➄❛➊✍➉➍➃❸➶❩➈⑧➆❋➏❛➉r➆★Ó☎➊☎➽❁➒❸➏❛➈✍➉✏➆✥➌❁➎→➏❛➂❁➈✍➌→➆ ➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð⑧➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➏❛➂➇➆❫➏Ò➏❛➋♠Ð❭➈✍➏❛➂❁➈❄➉✠➅✚➉❛➋❫➚♠➃❸➎✚➈☎Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎✞➄❛➈✍➉➍➓ ➚♠➃❸➊✍➈♠➻ ☎✠➄✠➎❤➃❸➄❛➊✍➐✚➄❛➄❛➈✍➎▼➶✚➉➍➃❸➈➡➱➇➘ ➃❸➌⑧➲❼➈❄➊✍➏➍➃❸➋♠➌✫✝❤➹✰➆✆➌✚➈✍➏➍➪✦➋♠➉❛✃✪➊r➒✇➃❸➈✍➌❤➏✯➔✂➐✚➄❛➏ ➊r➂❩➆✥➉r➆❫➊✍➏❛➈✍➉➍➃❸à❄➈✯➃❸➏❛➄✠➏❛➉r➆✍Ó☎➊✠➒❸➋❤➆✥➎Ñ➏❛➋✵➏❛➂✚➈✹➌✚➈✍➏➍➪✖➋❭➉❛✃❩➹☛➄❛➋✵➏❛➂➇➆❫➏✠➏❛➂✚➈✹➌✚➈✍➏➍➓ ➪✖➋❭➉❛✃❋➊✥➆❫➌▼➊❄➋♠➔✵➔✯➃❸➏✹➶➇➆❫➌✚➎✚➪Ô➃❸➎✚➏❛➂→➆❫➌✚➎▼➔☎➆✥➌❩➆✥Ð♠➈☎➷❤➐❁➈❄➐❁➈✍➄✹➃❸➌❞➆☎➪❳➆✍➘ ➏❛➂➇➆❫➏✹➉❛➈✥➆✍➒✇➃❸à✍➈✍➄✵➏❛➂✚➈✵➄❛➈✍➉❛➚♠➃❸➊✍➈☎➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈❄➌❤➏✍➻☎➼Ñ➈✵➐✚➄❛➈✏➆✆➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉ ➑❦➋❭➉❛➔❅➋❫➑Ò➏❛➉r➆★Ó➾➊✵➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌→➊✥➆★➒✇➒❸➈✍➎ ➆Ñ↕❦➫✥➧❤➙✥↔ ▼✥❮✚➤➡➧❤➙✥↕✧➽✚➒❸➏❛➈✍➉✍➻ ☎✽➏❛➋♠✃❭➈✍➌❋➶❁➐✚➊r✃♠➈✍➏✠➽✚➒❸➏❛➈✍➉➴➃❸➄✹➊r➂➇➆❫➉r➆❫➊❄➏❛➈✍➉➍➃❸à✍➈✍➎▼➶❤➘✏➏➍➪✖➋✲➅❩➆✥➉r➆❫➔✵➈✍➏❛➈❄➉❛➄✍➹❢➆ ➉r➆❫➏❛➈✡✠✪➆✥➌❁➎→➆❋➎❁➈✍➅✚➏❛➂☞☛★➻ ✫✠➌❁➈✏➊✥➆❫➌❞➏❛➂❤➃❸➌❁✃⑧➋❫➑✹➏❛➂✚➈☎➏❛➋♠✃❭➈✍➌❞➶❁➐✚➊➡✃❭➈✍➏ ➆❫➄✵➽✚➒✇➒✇➃❸➌❁Ð→➐❁➅→➪✖➃❸➏❛➂❑➏❛➋♠✃♠➈❄➌❁➄☎➊✍➋❭➌❤➏➍➃❸➌❤➐❁➋❭➐❁➄➍➒❸➘✷➆❫➏☎➆✞➉r➆❫➏❛➈✌✠♠➹❳➪✖➃❸➏❛➂✍☛ ➶❩➈➡➃❸➌✚Ð✆➃❸➏❛➄Ò➔☎➆❫➬♠➃❸➔➾➆★➒✦➎❁➈✍➅✚➏❛➂☛➻✏✎❢➚♠➈✍➉❛➘✏➏➍➃❸➔✵➈✲➆➾➅➇➆❫➊r✃♠➈❄➏➴➃❸➄✠Ð❭➈✍➌✚➈✍➉r➆❫➏❛➈✍➎ ✑ ➏❛➋♠✃❭➈✍➌✚➄✵➆✥➉❛➈✆➉❛➈❄➔✵➋✥➚♠➈✍➎Ñ➑s➉❛➋♠➔❅➏❛➂✚➈☎➶❁➐✚➊➡✃❭➈✍➏✍➹✿➪❳➂❁➈✍➉❛➈ ✑ ➃❸➄✵➏❛➂❁➈✲➄➍➃❸à✍➈ ➋❫➑Ô➏❛➂❁➈✵➅➇➆❫➊r✃♠➈❄➏✍➻✠☎✔➏❛➉r➆✍Ó☎➊✵➄❛➋❭➐❁➉❛➊✍➈☎➊✍➋❭➌❼➑s➋♠➉❛➔✵➄✹➏❛➋ ➆✆➏❛➋♠✃❭➈✍➌✞➶❁➐✚➊➡✃❭➈✍➏ ➽✚➒❸➏❛➈✍➉✆➭✒✠✔✓✄☛r➺➴➃✇➑❳➏❛➂✚➈✍➉❛➈☎➆❫➉❛➈➾➆★➒❸➪❳➆✍➘❤➄✂➈✍➌❁➋❭➐❁Ð❭➂▼➏❛➋♠✃❭➈✍➌✚➄P➃❸➌✞➏❛➂❁➈✲➶❁➐✚➊➡✃❭➈✍➏ ➪❳➂❁➈❄➌❁➈✍➚❭➈✍➉✹➆✂➅❩➆✥➊➡✃❭➈✍➏✖➃❸➄✠Ð♠➈❄➌❁➈✍➉r➆✥➏❛➈✍➎☛➻ â✪➋♠➉❛➈✖➅✚➉❛➈✍➊➡➃❸➄❛➈➡➒❸➘❭➹❁➊❄➋♠➌✚➄➍➃❸➎❁➈✍➉✦➆➴➅➇➆❫➊r✃♠➈❄➏❢Ð♠➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌☎➅❁➉❛➋❤➊✍➈❄➄❛➄✳➪✖➃❸➏❛➂ ✕✁✖ ➆✥➌❁➎ ✑ ✖ ➎❁➈✍➌✚➋♠➏➍➃❸➌✚Ð✂➏❛➂❁➈❳Ð♠➈❄➌❁➈✍➉r➆✥➏➍➃❸➋♠➌☎➏➍➃❸➔✵➈❳➆✥➌❁➎✲➄➍➃❸à✍➈♠➹❤➉❛➈✍➄❛➅➇➈✍➊✍➏➍➃❸➚♠➈r➒❸➘♠➹ ➋❫➑✿➏❛➂✚➈✘✗ ✳ ➏❛➂❋➅➇➆❫➊r✃♠➈✍➏✍➻❳➼✞➈✯➄r➆✍➘✏➏❛➂❩➆✥➏✠➏❛➂❤➃❸➄✹➏❛➉r➆★Ó☎➊✹➄❛➋❭➐❁➉❛➊✍➈✲➊✍➋♠➌❤➑❦➋♠➉❛➔✲➄ ➏❛➋⑧➆ ➏❛➋❭✃♠➈✍➌❞➶✚➐❁➊r✃♠➈✍➏✵➽✚➒❸➏❛➈✍➉✪➭✒✠✔✓✙☛➡➺✯➋✥➑Ò➉r➆❫➏❛➈✚✠❋➆✥➌❁➎→➎❁➈❄➅❁➏❛➂✛☛✲➃✇➑➮➏❛➂❁➈ ➄❛➈✍➷❤➐✚➈✍➌✚➊✍➈✢✜ ✖ ➎❁➈r➽❩➌✚➈✍➎❑➶❤➘✍✜✤✣✦✥✧☛✏➆❫➌✚➎✍✜ ✖ ✥✩★✍✪✬✫ ✍ ☛✭✓✄✜ ✖✯✮ å✱✰ ➭ ✕ ✖✳✲ ✕ ✖✴✮ å ➺✁✠ ✲ ✑ ✖ ✎❳➋♠➶➇➈✍➘❤➄☎➏❛➂❁➈✏➊✍➋♠➌✚➄❛➏❛➉r➆★➃❸➌❤➏☎➏❛➂➇➆❫➏✵✜ ✖✷✶ ✕✪➑❦➋♠➉➾➆★➒✇➒ ✗➍➻✵➁❳➂✚➈✵➷❤➐❩➆✥➌❤➏➍➃❸➏➍➃❸➈✍➄✵✜ ✖ ➹☛➃✇➑❳➌✚➋♠➌✚➌❁➈❄Ð❤➆❫➏➍➃❸➚❭➈♠➹✿➉❛➈✍➅✚➉❛➈✍➄❛➈✍➌❤➏✂➏❛➂❁➈✵➌❤➐✚➔✂➶❩➈✍➉ ➋❫➑✿➏❛➋❭✃♠➈✍➌✚➄❳➉❛➈✍➄➍➃❸➎❤➃❸➌❁Ð☎➃❸➌❋➏❛➂❁➈✹➶✚➐❁➊r✃♠➈❄➏✠➆✍➑❦➏❛➈✍➉✖➏❛➂✚➈✸✗✷✳ ➏❛➂✏➅❩➆✥➊➡✃❭➈✍➏✖➒❸➈✥➆✍➚♠➈✍➄✍➻ ✛❁➋❭➉✿➆➴Ð❫➃❸➚❭➈✍➌☎➏❛➉r➆★Ó☎➊✦Ð♠➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌☎➅❁➉❛➋❤➊✍➈❄➄❛➄✍➹❁➪✦➈✳➊✥➆✥➌✵➎✚➈➡➽❩➌✚➈❳➏❛➂❁➈❳➌✚➋♠➌❤➓ ➃❸➌✚➊✍➉❛➈✥➆❫➄➍➃❸➌✚Ð✪➑❦➐✚➌❁➊❄➏➍➃❸➋♠➌✦☛★➭✒✠✥➺✠➆✥➄❳➏❛➂❁➈✂➔✯➃❸➌❤➃❸➔☎➆★➒✦➚❭➆★➒❸➐✚➈✯➄❛➐✚➊➡➂❋➏❛➂➇➆❫➏Ò➏❛➂❁➈ ➅✚➉❛➋❤➊✍➈✍➄❛➄✠➊❄➋♠➌❤➑❦➋♠➉❛➔✵➄❳➏❛➋☎➆☎➭✒✠✭✓✙☛✍➭✒✠❫➺❛➺✳➽✚➒❸➏❛➈✍➉✍➻ ➯➍➌✹➉❛➈✍➊✍➈❄➌❤➏✻➘♠➈★➆❫➉❛➄✍➹✥➄❛➈✍➚❭➈✍➉r➆★➒✰↕❦➢✔✂➙✏▲➍➥❄↕◗➝✆✔✖Ï❳➶➇➆❫➄❛➈✍➎✲➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄✿➂❩➆✍➚♠➈ ➶❩➈✍➈❄➌✵➎❁➈❄➚♠➈➡➒❸➋❭➅❩➈✍➎☛➻✖➁❳➂❁➈❄➄❛➈✠➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✵➄❳➏r➆✥✃♠➈Ò➆❫➄✻➃❸➌✚➅❁➐✚➏✖➄❛➋❭➔✵➈❳➅❁➉❛➈✥➆✥➄➍➓ ➄➍➃❸Ð♠➌✚➈✍➎✞➆✥➅❁➅❩➋❭➉❛➏➍➃❸➋♠➌✚➔✵➈✍➌❤➏✹➋✥➑✿➏❛➂❁➈❳➒✇➃❸➌❁✃❋➈✍➬❤➅✚➉❛➈✍➄❛➄❛➈✍➎✞➆✥➄❳➆✯➄❛➈✍➏❳➋✥➑❢➉r➆❫➏❛➈❄➄ ✠✺✹✸➭❦➪➴➂✚➈✍➉❛➈✵✻Ý➒❷➆✥➶❩➈➡➒❸➄☎➏❛➂✚➈➾➱➇➋✥➪❳➄r➺❀✰❢➏❛➂❁➈☎➉❛➈✍➄❛➐❤➒❸➏➍➃❸➌✚Ð▼➎❁➈r➒❷➆★➘❤➄✲➎❁➈✍➅➇➈✍➌❁➎ ➋♠➌✏➏❛➂❁➈✂➶✚➐❁➊r✃♠➈✍➏❳➄➍➃❸à✍➈❄➄✏☛✄✹❱➭✒✠✺✹❩➺➡➻ ✫✠➌❁➈❳➋❫➑✰➏❛➂✚➈✖➽❩➉❛➄❛➏✿➄❛➐❁➊r➂✵➏➍➃❸➔✵➈r➓➍➄❛➏r➆✥➔✵➅✏➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄❳➪❳➆❫➄✿➏❛➂❁➈✌✡☛➝❤➢❷➜ ☞✖❮✚➙✥❮❁➙✥➢③↔ä➛✆➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➃❸➌❤➏❛➉❛➋❤➎✚➐✚➊✍➈✍➎❑➃❸➌✵✬✖➈➡➑❦➈✍➉❛➈❄➌❁➊✍➈✪✍✟✲✎❏➻→➁➴➂❤➃❸➄❋➆✍➒✇➓ Ð♠➋❭➉➍➃❸➏❛➂❁➔❐➪❳➆✥➄✹➏r➆❫➉❛Ð❭➈✍➏❛➈✍➎⑧➆❫➏✹➏❛➂✚➈✵➏❛➉r➆❫➎❤➃❸➏➍➃❸➋♠➌➇➆★➒Ò➎➇➆❫➏r➆✆➄❛➈❄➉❛➚♠➃❸➊✍➈✏➆❫➉❛➊r➂❼➃✇➓ ➏❛➈✍➊✍➏❛➐✚➉❛➈♠➹❱➆❫➌✚➎▼➄❛➋✆➃❸➌❤➚❭➋❫➒❸➚❭➈✍➎❞➌❁➋✏➅❁➉❛➈✥➆✍➒✇➒❸➋❤➊✥➆❫➏➍➃❸➋❭➌✸➋✥➑❳➉❛➈✍➄❛➋♠➐✚➉❛➊✍➈✍➄☎➭❏➆❫➌✚➎ ➏❛➂❤➐✚➄❋➂❩➆✥➎⑤➈✥➆✥➊➡➂✼✠ ✹ ✥✾✽⑤➪❳➂✚➈✍➉❛➈✦✽⑤➎❁➈❄➌❁➋❭➏❛➈✍➄❋➏❛➂❁➈✞➒✇➃❸➌❁✃✫➄❛➅❩➈✍➈✍➎➇➺➡➻
n addition, a weighted version of the Fair Queueing algo flow long the path would be at least r rithm (which we refer to as WFQ), in which the r"neee Thus, all of the eing delay would occur in the leaky lot all be equal, was also briefly described in Reference [fB bucket filter and the flow obeys an(r,b(r)) toke The Virtuasack algorit hm, described in Refere e6B bucket filter, the delay in the leaky bucket filter would be olves an extremely similar underlying packet scheduling bounded by b(r)/r. Notice that the delay bound of algorithm, but was expressly designed for a context where icular flow is indo dent of the other fe istics; they can be arbitrarily badly behaved and the bound tal part of its architecture the assumption that the shares r still applies. Furthermore, these bounds are strict, in th were arbitrary. Parekh and Pallager in Reference [QB rein- they can be realized with a set of greedy sources which keep troduce the WfQ algorithm under the name of packetized their token buckets empty nera=zed processor sharing(PPPS). They have proven the The previous paragraphs describe WFQ in the fluid flo approx et ized version of the ions, can deliver a guaranteed quality of service([2tB. We algorit hm in a straightforward manner. Define 8;(t) for all resent a brief summary of the wFQ algorithm below, sinc t>t as the number of bits that haye been serviced from we make use of it in our overall scheduling algorit hm; see the flow a bet ween the times ti and t. g ssociate with each References [, 2t Bfor more details packet the function E;()=(m(9)j 8:())/r where we of fe f c=ack rates take the right-hand limit of t his number is the level of The clock rate of a flow represents the relative share of backlog ahead of the packet i in the flow as queue divided presents the pro portion of the total link bandwidth which expected delay until depart ure for the last bit in the packe,? the link bandwidth this flow is entitled to; more properly, it by the flow's share of the link, and can be thought of this flow will receive when it is active. By assigning it a The packetized version of WFQ is merely, at any time,o clock rate ra the network commits to provide to this flow Then the next packet to be transmit ted must be chosen an effective throughput rate no worse than(ur)/Car) select the packet with the minimal E?(t). This algorithm is called a time-st amp based scheme because there is an al where the sum in the denominator is over all currently active terna tive but equivalent formulation in which each packet is This formulation can be made precise in the context of a stamped with a time-stamp as it arrives and th luid flow model of the net work, where the bits drain contin- are transmit ted in increasing order of time-stamps; see Ref- uously out of the queue. Let t? and pa denote the generation erences [, 2t Bfor details on this formulation. Parekh and ath flow. We define the set of functions m"(), which char- bound, applies to this packet ized algorit hm as well. Hor acterize at any time the backlog of bits which each source ever, the formulae for the de lays in the packe tized case are and set m (t)=t. We say that a flow is active significantly more complicated; see Re ference [2t Bor details To underst and the relation bet ween the clock rate r, the Then the dynamics of the system are determined as follows. bucket size b(r)and the resultant delay, consider what hap- Whenever a packet arrives, m must discont inuously increase pens to a burst of packets. The packet that receives the y the packet highest queueing delay is the last packet of a burst. The n(tt)and m(t-)refer to right hand and left hand limits bound on the jit ter is proport ional to the size of the burst of m at t. gt all other times, we know that the bits are and inversely proportional to the clock ra te. The means by draining out of the queues of the active flows in proportion to the clock rates of the respective fow increase its clock rate r to permit the burst to pass through This comp mo ra laEA(, amo(=t if a E A( worst-case bounds (i.e. the bounds must apply fo ble behaviors of the ot her sources ) the primary function of This completely characterizes the dynamics of the fluid a scheduling algorit hm designed to deliver guaranteed flow model. Parekh and P allager have shown the remark is to isoamte fows from each ot he that a f able result that, in a network with arbitrary topology, if a have only a limited negative e ffect on other flows. The WFQ flow gets the same clock rate at every switch and the sum of scheme isolates each source from the others by providing it a the clock rates of all the flows at every switch is no greater specified share of the bandwid th under overload conditions. than the link speed, then the queueing delay of that flow is The work of Parekh and P allager provides a way for tI bounded above by b°(r Intuitively this bound is the source to compute the maximum queueing delay which its delay that would result from an inst ant aneous packet burst packets will encounter, provided that the source rest ricts of the token bucket size being serviced by a single link of self to a(, b)token bucket filter. But the network's schedul rate r; the queueing delays are no worse than if the ent ir ing algorithm does not depend on this filter. Indeed, an im. by noting that if flow a's traffic were put through a leaky P iven that the wfQ algorithm can deliver guaranteed bucket filter of rate r at the edge of the net work, then the service, why not use this algorithm to provide predicted ser flow would not suffer any furt her queueing delays within the vice as well? Recall that the goal of guaranteed service was network since the instantaneous service rate given to this to provide worst-case guarantees, which are required by in- 5 The we_rhted of tar 2_ng s ment oned on tolerant and rigid applications. The goal of predicted service is not to provide worst-case delay bounds, but rat her to min the actual post facto delay bounds the out at a jud jow ' fron of a saky bucket of rate', the bis dr ive applications to reel in the point. In the next section, we show that the wfQ algo-
➯❏➌❞➆❫➎✚➎❤➃❸➏➍➃❸➋♠➌☛➹❳➆✆➪✦➈➡➃❸Ð♠➂❤➏❛➈✍➎❞➚❭➈✍➉❛➄➍➃❸➋♠➌❞➋❫➑Ô➏❛➂❁➈✆✛➇➆★➃❸➉ ✂✠➐❁➈✍➐✚➈➡➃❸➌✚Ð▼➆✍➒❸Ð♠➋✥➓ ➉➍➃❸➏❛➂✚➔❰➭❦➪❳➂❼➃❸➊r➂❞➪✖➈✪➉❛➈r➑❦➈✍➉✵➏❛➋✞➆✥➄✵➼✛✂✹➺➡➹ ➃❸➌→➪❳➂❼➃❸➊r➂❞➏❛➂❁➈✡✠ ✹✸➌✚➈✍➈✍➎ ➌✚➋♠➏Ô➆★➒✇➒☛➶❩➈Ò➈✍➷❤➐➇➆★➒❦➹❩➪Ô➆❫➄✦➆★➒❸➄❛➋✂➶❁➉➍➃❸➈➡➱➇➘☎➎❁➈❄➄❛➊✍➉➍➃❸➶❩➈✍➎☎➃❸➌ ✬✖➈➡➑s➈✍➉❛➈✍➌✚➊✍➈☛✍✟✲✎ ➻ ➁❳➂✚➈✂✁❱➢③➜r↕◗❮✚➝✖✥✓✥❸➫✥➤r➧✻➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✪➹✚➎✚➈✍➄❛➊✍➉➍➃❸➶❩➈✍➎✂➃❸➌❉✬Ô➈r➑❦➈✍➉❛➈✍➌✚➊✍➈✍➄❏✍ ✮✖✗❤➹❏✮✖✙✭✎❏➹ ➃❸➌❤➚❭➋❫➒❸➚❭➈✍➄☎➆❫➌✞➈✍➬❤➏❛➉❛➈✍➔✲➈➡➒❸➘Ñ➄➍➃❸➔✂➃✇➒❷➆❫➉✵➐✚➌❁➎✚➈✍➉➍➒❸➘♠➃❸➌✚ÐÑ➅➇➆❫➊r✃♠➈❄➏✹➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð ➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✪➹✦➶❁➐✚➏✂➪❳➆❫➄✹➈✍➬❤➅✚➉❛➈✍➄❛➄➍➒❸➘▼➎✚➈✍➄➍➃❸Ð❭➌❁➈✍➎❞➑❦➋♠➉✂➆☎➊✍➋♠➌❤➏❛➈❄➬❤➏✯➪❳➂✚➈✍➉❛➈ ➉❛➈❄➄❛➋♠➐✚➉❛➊✍➈✍➄✖➪✖➈❄➉❛➈➴➅✚➉❛➈✥➆✥➅❁➅❩➋❭➉❛➏➍➃❸➋♠➌✚➈✍➎✞➆❫➌✚➎✵➏❛➂❤➐✚➄✖➂➇➆❫➎☎➆✥➄✖➆➴➑❦➐✚➌✚➎❩➆✥➔✵➈✍➌❤➓ ➏r➆✍➒❩➅➇➆❫➉❛➏❢➋✥➑❩➃❸➏❛➄✦➆❫➉❛➊r➂❼➃❸➏❛➈✍➊❄➏❛➐❁➉❛➈❳➏❛➂✚➈❳➆❫➄❛➄❛➐✚➔✵➅✚➏➍➃❸➋♠➌✪➏❛➂➇➆❫➏❢➏❛➂✚➈✖➄❛➂➇➆❫➉❛➈✍➄ ✠ ✹ ➪✦➈✍➉❛➈❳➆❫➉❛➶❤➃❸➏❛➉r➆❫➉❛➘❭➻❳➵❢➆❫➉❛➈❄✃❤➂☎➆❫➌✚➎✆✑✹➆★➒✇➒❷➆✥Ð♠➈✍➉✍➹✚➃❸➌ ✬✖➈➡➑❦➈✍➉❛➈❄➌❁➊✍➈✌✍✒✂✔✓✏✎➍➹❤➉❛➈r➃❸➌❼➓ ➏❛➉❛➋❤➎✚➐✚➊✍➈✪➏❛➂✚➈☎➼✛✂ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➐✚➌❁➎✚➈✍➉✵➏❛➂✚➈☎➌❩➆✥➔✵➈☎➋❫➑❳Ï➇➝❤➤r➧❼➙✥↕❦➢☎✄✥➙➡➞ ➛❤➙✥↔❩➙✥➜➍➝✖✥❸➢✄❫➙r➞✻Ï✰➜➍➫✥➤➡➙✍➥r➥✍➫❤➜✻➥■✦❩➝❤➜r➢③↔❤➛✳➭❏➵❋✑✹➵✳➲✚➺➡➻♠➁❳➂✚➈✍➘✹➂➇➆★➚❭➈✖➅✚➉❛➋❫➚❭➈✍➌✵➏❛➂✚➈ ➃❸➔✲➅❩➋♠➉❛➏r➆✥➌❤➏✆➉❛➈❄➄❛➐❼➒❸➏✵➏❛➂➇➆❫➏✲➏❛➂❼➃❸➄☎➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✪➹❳➐✚➌✚➎❁➈✍➉☎➊✍➈❄➉❛➏r➆★➃❸➌❞➊✍➋♠➌✚➎❤➃✇➓ ➏➍➃❸➋❭➌❁➄✍➹☛➊★➆❫➌✪➎✚➈➡➒✇➃❸➚❭➈✍➉✂➆✯Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❋➷❤➐➇➆★➒✇➃❸➏❏➘❋➋❫➑✿➄❛➈✍➉❛➚♠➃❸➊✍➈☎➭✢✍ ✮✖✕✭✎s➺➡➻❢➼✞➈ ➅✚➉❛➈✍➄❛➈✍➌❤➏Ô➆✠➶❁➉➍➃❸➈➡➑✰➄❛➐✚➔✵➔☎➆✥➉❛➘✵➋❫➑✰➏❛➂✚➈➴➼✛✂✫➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✾➶❩➈➡➒❸➋✥➪✹➹❁➄➍➃❸➌✚➊✍➈ ➪✦➈✆➔➾➆❫✃♠➈☎➐✚➄❛➈✪➋✥➑➴➃❸➏✹➃❸➌❞➋♠➐✚➉✵➋✥➚♠➈✍➉r➆✍➒✇➒Ò➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð→➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔❅✰❳➄❛➈✍➈ ✬✖➈➡➑s➈✍➉❛➈✍➌✚➊✍➈✍➄☛✍✟❁➹ ✮✬✕✲✎✧➑s➋♠➉❳➔✵➋♠➉❛➈✹➎✚➈✍➏r➆✍➃✇➒❸➄✍➻ ✛✧➃❸➉❛➄❛➏✍➹☛➊✍➋❭➌❁➄➍➃❸➎✚➈✍➉✂➄❛➋♠➔✵➈✹➄❛➈✍➏Ò➋❫➑❢➱➇➋✥➪➴➄➮➆❫➌✚➎✞➆✵➄❛➈✍➏✠➋✥➑Ò➤✭✥❸➫✥➤r➧☎➜➍➝❤↕❦➙✍➥ ✠ ✹ ➻✿➁❳➂❁➈✹➊➡➒❸➋❤➊r✃✪➉r➆✥➏❛➈✹➋❫➑✦➆P➱➇➋✥➪✛➉❛➈✍➅✚➉❛➈✍➄❛➈✍➌❤➏❛➄Ò➏❛➂❁➈✹➉❛➈➡➒❷➆✥➏➍➃❸➚♠➈✵➄❛➂➇➆❫➉❛➈✹➋✥➑ ➏❛➂✚➈➴➒✇➃❸➌✚✃✪➶➇➆❫➌✚➎✚➪Ô➃❸➎✚➏❛➂✪➏❛➂❤➃❸➄✖➱❩➋✥➪✸➃❸➄✖➈❄➌❤➏➍➃❸➏➍➒❸➈✍➎❋➏❛➋ ✰✚➔✵➋❭➉❛➈✠➅✚➉❛➋♠➅❩➈✍➉➍➒❸➘❭➹❁➃❸➏ ➉❛➈❄➅❁➉❛➈✍➄❛➈❄➌❤➏❛➄➴➏❛➂✚➈✠➅✚➉❛➋♠➅➇➋♠➉❛➏➍➃❸➋♠➌❋➋✥➑❢➏❛➂✚➈✠➏❛➋❭➏r➆★➒☛➒✇➃❸➌❁✃☎➶➇➆❫➌✚➎❁➪✖➃❸➎✚➏❛➂❋➪❳➂❼➃❸➊r➂ ➏❛➂❤➃❸➄✵➱❩➋✥➪ ➪✖➃✇➒✇➒✹➉❛➈✍➊✍➈r➃❸➚♠➈❋➪❳➂❁➈❄➌Ý➃❸➏✵➃❸➄✏➆❫➊✍➏➍➃❸➚❭➈♠➻ ✎✳➘❞➆✥➄❛➄➍➃❸Ð♠➌❤➃❸➌❁Ð❞➃❸➏✏➆ ➊r➒❸➋❤➊➡✃▼➉r➆✥➏❛➈ ✠ ✹ ➏❛➂✚➈☎➌❁➈✍➏❏➪✖➋♠➉❛✃✞➊✍➋❭➔✵➔✯➃❸➏❛➄✵➏❛➋✏➅❁➉❛➋✥➚♠➃❸➎✚➈❋➏❛➋❋➏❛➂❼➃❸➄✹➱❩➋✥➪ ➆✥➌Ñ➈rÞ☛➈✍➊✍➏➍➃❸➚❭➈✆➏❛➂✚➉❛➋♠➐✚Ð♠➂✚➅✚➐❁➏✂➉r➆❫➏❛➈✵➌✚➋✪➪✦➋♠➉❛➄❛➈✵➏❛➂➇➆❫➌❞➭✒✽ ✠ ✹➇➺✝✆❼➭✟✞✡✠ ✠ ✠ ➺ ➪❳➂✚➈✍➉❛➈✖➏❛➂✚➈✖➄❛➐✚➔ã➃❸➌✵➏❛➂✚➈✖➎✚➈✍➌✚➋♠➔✯➃❸➌➇➆❫➏❛➋❭➉ ➃❸➄❢➋✥➚♠➈❄➉✿➆★➒✇➒➇➊✍➐❁➉❛➉❛➈❄➌❤➏➍➒❸➘✵➆✥➊✍➏➍➃❸➚♠➈ ➱➇➋✥➪❳➄✍➻ ➁❳➂❼➃❸➄✿➑❦➋♠➉❛➔✂➐❼➒❷➆✥➏➍➃❸➋♠➌❋➊✥➆✥➌✵➶❩➈Ò➔☎➆❫➎✚➈✠➅✚➉❛➈✍➊➡➃❸➄❛➈❳➃❸➌✪➏❛➂✚➈➴➊❄➋♠➌❤➏❛➈✍➬❤➏✖➋✥➑❱➆ ➱➇➐❤➃❸➎✆➱➇➋❫➪❞➔✵➋❤➎✚➈➡➒☛➋✥➑✰➏❛➂❁➈❳➌✚➈✍➏➍➪✖➋❭➉❛✃❩➹❤➪❳➂❁➈❄➉❛➈➴➏❛➂✚➈❳➶❼➃❸➏❛➄✦➎❁➉r➆✍➃❸➌✆➊❄➋♠➌❤➏➍➃❸➌❤➓ ➐✚➋♠➐✚➄➍➒❸➘✂➋♠➐✚➏❢➋❫➑➇➏❛➂❁➈✿➷❤➐❁➈❄➐❁➈♠➻★✳☛➈✍➏ ✕ ✹✖ ➆✥➌❁➎ ✑ ✹✖ ➎❁➈❄➌❁➋❭➏❛➈✖➏❛➂❁➈✿Ð♠➈❄➌❁➈✍➉r➆✥➏➍➃❸➋♠➌ ➏➍➃❸➔✲➈✹➆❫➌✚➎✪➄➍➃❸à❄➈♠➹✚➉❛➈✍➄❛➅❩➈✍➊✍➏➍➃❸➚❭➈➡➒❸➘♠➹☛➋✥➑❢➏❛➂✚➈✏✗✷✳ ➏❛➂✲➅❩➆✥➊➡✃❭➈✍➏Ò➆❫➉❛➉➍➃❸➚♠➃❸➌✚Ð➾➃❸➌☎➏❛➂✚➈ ✻ ✳ ➏❛➂✯➱➇➋✥➪P➻✿➼✞➈❳➎❁➈➡➽➇➌✚➈✹➏❛➂❁➈Ò➄❛➈✍➏✳➋✥➑☛➑❦➐❁➌✚➊✍➏➍➃❸➋❭➌❁➄☞☛✹ ➭ ✕ ➺➡➹❤➪❳➂❼➃❸➊r➂☎➊➡➂➇➆❫➉➍➓ ➆✥➊✍➏❛➈✍➉➍➃❸à✍➈✏➆✥➏✵➆✥➌❤➘Ñ➏➍➃❸➔✵➈☎➏❛➂✚➈✵➶➇➆❫➊r✃♠➒❸➋♠Ð✞➋❫➑❳➶❤➃❸➏❛➄✵➪❳➂❼➃❸➊r➂▼➈✥➆✥➊➡➂⑧➄❛➋♠➐✚➉❛➊✍➈ ➂➇➆❫➄✦➏❛➋✹➄❛➈✍➌❁➎☛➹➇➆❫➌✚➎☎➄❛➈✍➏✌☛✹ ➭ ✕♠➺ ✥ ✕❼➻✿➼✞➈❳➄r➆★➘✲➏❛➂❩➆✥➏✖➆➴➱➇➋✥➪Ý➃❸➄❳➆✥➊✍➏➍➃❸➚♠➈ ➆✥➏❢➏➍➃❸➔✵➈ ✕ ➃✇➑✍☛✹ ➭ ✕ ➺✌✎✚✕ ✰❭➒❸➈✍➏✑✏✂➭ ✕ ➺✻➎✚➈✍➌✚➋♠➏❛➈❳➏❛➂❁➈✦➄❛➈✍➏❢➋❫➑✰➆✥➊✍➏➍➃❸➚♠➈✖➱➇➋✥➪➴➄✍➻ ➁❳➂✚➈✍➌☎➏❛➂❁➈Ò➎❁➘❤➌➇➆❫➔✂➃❸➊✍➄❳➋❫➑✰➏❛➂❁➈Ò➄❛➘❤➄❛➏❛➈✍➔✾➆✥➉❛➈✠➎✚➈✍➏❛➈✍➉❛➔✯➃❸➌✚➈✍➎ ➆❫➄❢➑❦➋❫➒✇➒❸➋✥➪❳➄✍➻ ➼→➂❁➈❄➌❁➈✍➚❭➈✍➉❳➆✠➅➇➆❫➊r✃♠➈✍➏✦➆❫➉❛➉➍➃❸➚❭➈✍➄✍➹✒☛ ➔✂➐❁➄❛➏✿➎❤➃❸➄❛➊✍➋♠➌❤➏➍➃❸➌❤➐✚➋♠➐✚➄➍➒③➘☎➃❸➌❁➊❄➉❛➈✥➆❫➄❛➈ ➶❤➘❋➏❛➂✚➈✂➅❩➆✥➊➡✃❭➈✍➏✠➄➍➃❸à✍➈❭❒✓☛✹ ➭ ✕✝✔ ➺✟✥✕☛✹ ➭ ✕ ✮ ➺ ✰ ✑ ✖ ➃✇➑ ✕ ✥ ✕ ✹✖ ➹☛➪❳➂✚➈✍➉❛➈ ☛✹✧➭ ✕ ✔ ➺✦➆❫➌✚➎✖☛✹✧➭ ✕ ✮ ➺✳➉❛➈➡➑s➈✍➉✳➏❛➋✂➉➍➃❸Ð♠➂❤➏❳➂➇➆❫➌✚➎❋➆✥➌❁➎✵➒❸➈➡➑❦➏✦➂❩➆✥➌❁➎☎➒✇➃❸➔✯➃❸➏❛➄ ➋✥➑☞☛✹ ➆✥➏ ✕ ➻✆☎➴➏✂➆✍➒✇➒❳➋♠➏❛➂✚➈✍➉✂➏➍➃❸➔✵➈✍➄✍➹✿➪✖➈✵✃❤➌✚➋✥➪✔➏❛➂➇➆❫➏✂➏❛➂❁➈☎➶❤➃❸➏❛➄✵➆✥➉❛➈ ➎✚➉r➆★➃❸➌❤➃❸➌✚Ð❋➋♠➐✚➏➴➋✥➑✿➏❛➂❁➈✂➷❤➐✚➈✍➐❁➈❄➄✠➋❫➑✿➏❛➂✚➈✲➆❫➊✍➏➍➃❸➚❭➈P➱➇➋✥➪➴➄✖➃❸➌❋➅✚➉❛➋♠➅❩➋❭➉❛➏➍➃❸➋♠➌ ➏❛➋✲➏❛➂❁➈✹➊➡➒❸➋❤➊r✃✪➉r➆✥➏❛➈✍➄❳➋❫➑✿➏❛➂✚➈✹➉❛➈✍➄❛➅❩➈✍➊❄➏➍➃❸➚♠➈✯➱➇➋✥➪➴➄✍❒ ✗✒☛✹ ➭ ✕ ➺ ✗✕ ✥ ✽ ✠ ✹ ✞✠✙✘✙✚✜✛✣✢✥✤ ✠ ✠ ➃✇➑ ✻✧✦★✏✂➭ ✕ ➺ ✓ ✗✜☛✹ ➭ ✕ ➺ ✗✕ ✥ ✕P➃✇➑ ✻✪✩✦✫✏✯➭ ✕ ➺ ➁❳➂❼➃❸➄✂➊✍➋❭➔✵➅❼➒❸➈❄➏❛➈➡➒❸➘Ñ➊r➂➇➆❫➉r➆❫➊❄➏❛➈✍➉➍➃❸à✍➈✍➄✂➏❛➂❁➈✲➎❁➘❤➌➇➆❫➔✯➃❸➊❄➄✵➋❫➑✦➏❛➂❁➈✂➱❩➐❤➃❸➎ ➱➇➋✥➪✔➔✵➋❤➎✚➈➡➒❦➻ ➵❢➆❫➉❛➈✍✃❤➂→➆✥➌❁➎✫✑✹➆✍➒✇➒❷➆❫Ð❭➈✍➉✵➂➇➆★➚❭➈☎➄❛➂❁➋✥➪❳➌▼➏❛➂✚➈✆➉❛➈❄➔☎➆❫➉❛✃♠➓ ➆✥➶❼➒❸➈☎➉❛➈✍➄❛➐❤➒❸➏✂➏❛➂❩➆✥➏✍➹✧➃❸➌❞➆☎➌❁➈❄➏➍➪✖➋❭➉❛✃❋➪Ô➃❸➏❛➂①➆❫➉❛➶❤➃❸➏❛➉r➆❫➉❛➘✞➏❛➋♠➅❩➋✥➒❸➋♠Ð♠➘❭➹ ➃✇➑➴➆ ➱➇➋✥➪✸Ð❭➈✍➏❛➄✖➏❛➂✚➈➴➄r➆✥➔✵➈✠➊r➒❸➋❤➊➡✃✵➉r➆✥➏❛➈Ò➆❫➏✿➈✍➚♠➈✍➉❛➘✲➄❛➪Ô➃❸➏❛➊r➂✏➆❫➌✚➎☎➏❛➂❁➈❳➄❛➐✚➔ ➋✥➑ ➏❛➂✚➈✂➊➡➒❸➋❤➊➡✃✏➉r➆❫➏❛➈✍➄Ò➋❫➑✦➆✍➒✇➒❢➏❛➂✚➈P➱➇➋✥➪➴➄Ò➆✥➏✠➈✍➚❭➈✍➉❛➘☎➄❛➪Ô➃❸➏❛➊r➂✪➃❸➄✠➌✚➋☎Ð♠➉❛➈✥➆✥➏❛➈✍➉ ➏❛➂➇➆❫➌❋➏❛➂✚➈✠➒✇➃❸➌✚✃❋➄❛➅❩➈✍➈✍➎☛➹❩➏❛➂✚➈✍➌❋➏❛➂✚➈✹➷❤➐❁➈✍➐✚➈➡➃❸➌✚Ð✪➎✚➈➡➒❷➆✍➘❋➋❫➑❱➏❛➂❩➆✥➏➴➱➇➋❫➪❑➃❸➄ ➶➇➋♠➐✚➌❁➎✚➈✍➎⑧➆❫➶❩➋✥➚♠➈✂➶❤➘✚☛✄✹✧➭✒✠✺✹☛➺✝✆ ✠✺✹✧➻✿➯➍➌❤➏❛➐❤➃❸➏➍➃❸➚♠➈r➒❸➘♠➹❢➏❛➂❤➃❸➄✠➶❩➋❭➐❁➌✚➎✪➃❸➄✠➏❛➂✚➈ ➎✚➈➡➒❷➆✍➘✪➏❛➂➇➆❫➏❳➪✖➋❭➐❼➒❸➎❋➉❛➈❄➄❛➐❼➒❸➏❳➑❦➉❛➋♠➔ ➆✥➌✆➃❸➌✚➄❛➏r➆❫➌❤➏r➆❫➌✚➈✍➋❭➐❁➄✹➅➇➆❫➊r✃♠➈✍➏Ò➶❁➐✚➉❛➄❛➏ ➋✥➑Ò➏❛➂❁➈✏➏❛➋♠✃♠➈❄➌Ý➶✚➐✚➊➡✃❭➈✍➏✵➄➍➃❸à✍➈✏➶❩➈➡➃❸➌✚Ð▼➄❛➈✍➉❛➚♠➃❸➊✍➈✍➎→➶❤➘❞➆✪➄➍➃❸➌✚Ð❫➒❸➈✪➒✇➃❸➌✚✃Ý➋✥➑ ➉r➆✥➏❛➈✸✠ ✹ ✰✚➏❛➂✚➈✯➷❤➐✚➈✍➐✚➈➡➃❸➌✚Ð✪➎✚➈➡➒❷➆✍➘❤➄✹➆❫➉❛➈✂➌✚➋✵➪✖➋♠➉❛➄❛➈✹➏❛➂➇➆❫➌✪➃✇➑✿➏❛➂✚➈✹➈✍➌❤➏➍➃❸➉❛➈ ➌✚➈✍➏❏➪✖➋♠➉❛✃✞➪✖➈✍➉❛➈☎➉❛➈✍➅❤➒❷➆❫➊✍➈❄➎Ý➶❤➘⑧➆✪➄➍➃❸➌✚Ð❫➒❸➈✪➒✇➃❸➌✚✃▼➪✖➃❸➏❛➂❞➆✪➄❛➅❩➈✍➈✍➎❞➈✍➷❤➐➇➆★➒ ➏❛➋✞➏❛➂✚➈✆➱➇➋❫➪✗✳ ➄✵➊➡➒❸➋❤➊r✃❞➉r➆❫➏❛➈✚✠ ✹ ➻❞➁❳➂❼➃❸➄✲➉❛➈✍➄❛➐❼➒❸➏☎➊✥➆✥➌✸➶❩➈✏➔✵➋♠➏➍➃❸➚❭➆✥➏❛➈✍➎ ➶❤➘✞➌✚➋♠➏➍➃❸➌✚ÐÑ➏❛➂➇➆❫➏✹➃✇➑✖➱➇➋❫➪ ✻ ✳ ➄Ò➏❛➉r➆★Ó☎➊☎➪✖➈❄➉❛➈✵➅❁➐✚➏✂➏❛➂❁➉❛➋❭➐❁Ð❭➂❞➆✆➒❸➈✥➆❫✃❤➘ ➶✚➐✚➊➡✃❭➈✍➏✳➽✚➒❸➏❛➈✍➉✖➋❫➑✰➉r➆✥➏❛➈✏✠✺✹☎➆❫➏✖➏❛➂✚➈❳➈✍➎❁Ð❭➈✠➋✥➑✻➏❛➂✚➈❳➌❁➈❄➏➍➪✖➋❭➉❛✃✭✬❼➹ä➏❛➂❁➈✍➌☎➏❛➂✚➈ ➱➇➋✥➪✸➪✖➋❭➐❼➒❸➎☎➌✚➋♠➏✿➄❛➐❼Þ❙➈✍➉❳➆❫➌❤➘✹➑❦➐✚➉❛➏❛➂❁➈❄➉✳➷❤➐✚➈✍➐✚➈➡➃❸➌❁Ð✲➎❁➈➡➒❷➆✍➘❤➄❳➪✖➃❸➏❛➂❼➃❸➌☎➏❛➂✚➈ ➌✚➈✍➏❏➪✖➋♠➉❛✃❞➄➍➃❸➌✚➊✍➈❋➏❛➂❁➈☎➃❸➌❁➄❛➏r➆✥➌❤➏r➆❫➌✚➈✍➋♠➐✚➄❋➄❛➈✍➉❛➚♠➃❸➊✍➈❋➉r➆✥➏❛➈❋Ð❫➃❸➚❭➈✍➌✸➏❛➋✞➏❛➂❼➃❸➄ ❫ð✰ì✥ç✖ñ☛ç✔✘ û✍ì✍í❦ç❛õ✗✙rç❛ê❦è▲✘ ô✍ÿ✵ô★ ✡♠é✞✘ ê✯✮✻ò❫ç❛ò✥ç✔✘ÿ❫û✗✘ è ✚✦ç❛ÿ✍í✆✘ ô✍ÿ✥ç❛õ✂ô✍ÿ✯ó❭é➡û❄ç ❀✓ ô ★ æ✰ç★ç❛ê❦ç❛ÿ❫ë❛ç ✘✓✚ ✠ í❦ì✥ô✍ò✥û✍ì✯ÿ✥ô✍í✿ê❦ç★ç❛ê❦ê◗ç❛õ✂í❦ô✹ö✍÷✂í◗ì✥ç❳ÿ❭é❀✚✖ç✱✰✹ç✔✘ û❄ì✍í❦ç❛õ ✡❭é❀✘ ê ✮✻ò✥ç❛ò✥ç✔✘ ÿ✥û✒✪ ✬ ï❷ÿ✂é✯✲äò✟✘ õ✓✲❤ôrñ❋✙➡ç❛ê❦è✆✘ ô✍ÿ✂ô★ é❂✩ çré❬✍÷✠ö✥ò✥ë❬rç❛í✿ô★ ê❏é➡í❦ç✌✳ ✠ í◗ì✥ç✖ö✟✘ í◗è✿õ✥ê❏é❀✘ ÿ ô❄ò❫í❱é➡í❢é✖ë❛ô✍ÿ✥è❦í❏é➡ÿ✍í✰ê➍érí◗ç✴✳Ôé➡ÿ✥õ✂é➡ÿ✍÷❳ç✔❊✍ë❛ç❛è❦è✼✘ è✶✵✍ò✥ç❛ò❫ç❛õ✻✪ ➱➇➋❫➪ ➆✥➏✯➈❄➚♠➈✍➉❛➘▼➄❛➪✖➃❸➏❛➊r➂→➆★➒❸➋♠➌✚Ð▼➏❛➂✚➈✪➅➇➆❫➏❛➂❞➪✖➋❭➐❼➒❸➎❞➶❩➈ ➆❫➏✯➒❸➈✥➆✥➄❛➏✵✠ ✹ ➻ ➁❳➂❤➐❁➄✍➹Ò➆✍➒✇➒✹➋✥➑✂➏❛➂❁➈❋➷❤➐✚➈✍➐✚➈➡➃❸➌❁Ð→➎❁➈r➒❷➆★➘→➪✖➋♠➐❤➒❸➎❑➋❤➊✍➊✍➐✚➉✵➃❸➌❑➏❛➂❁➈✪➒❸➈★➆❫✃❤➘ ➶✚➐❁➊r✃♠➈✍➏✖➽✚➒❸➏❛➈✍➉Ò➆❫➌✚➎✧➹➇➄➍➃❸➌❁➊❄➈✯➏❛➂✚➈➴➱➇➋✥➪⑤➋❭➶❩➈✍➘❤➄Ò➆❫➌✞➭✒✠ ✹ ✓✄☛ ✹✻➭✒✠✺✹❩➺❛➺✖➏❛➋♠✃❭➈✍➌ ➶✚➐❁➊r✃♠➈✍➏✹➽✚➒❸➏❛➈✍➉✍➹✿➏❛➂❁➈☎➎✚➈➡➒❷➆✍➘Ñ➃❸➌❞➏❛➂❁➈✯➒❸➈★➆❫✃❤➘▼➶✚➐❁➊r✃♠➈❄➏✯➽❁➒❸➏❛➈❄➉✯➪✖➋❭➐❼➒❸➎❞➶❩➈ ➶❩➋❭➐❁➌✚➎❁➈❄➎✸➶❤➘ ☛✄✹✻➭✒✠ ✹➇➺✝✆ ✠✺✹✻➻▼➸❳➋♠➏➍➃❸➊✍➈✏➏❛➂❩➆✥➏✵➏❛➂✚➈✆➎✚➈➡➒❷➆✍➘❞➶❩➋♠➐✚➌✚➎✸➋❫➑➮➆ ➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✹➱❩➋✥➪⑤➃❸➄✲➢③↔❩➞❤➙➍Ï➇➙✥↔❩➞❤➙✍↔❩↕❤➋❫➑✿➏❛➂✚➈✂➋♠➏❛➂✚➈✍➉➴➱➇➋✥➪➴➄☛✳☛➊r➂➇➆❫➉r➆❫➊❄➏❛➈✍➉➍➓ ➃❸➄❛➏➍➃❸➊✍➄☛✰☛➏❛➂✚➈✍➘✵➊✥➆✥➌✪➶❩➈✂➆✥➉❛➶❼➃❸➏❛➉r➆✥➉➍➃✇➒❸➘❋➶❩➆✥➎❼➒❸➘✏➶❩➈✍➂➇➆✍➚♠➈✍➎ ➆❫➌✚➎✪➏❛➂✚➈✹➶❩➋♠➐✚➌❁➎ ➄❛➏➍➃✇➒✇➒Ò➆✥➅❁➅❤➒✇➃❸➈✍➄✍➻✞✛❁➐✚➉❛➏❛➂✚➈✍➉❛➔✵➋♠➉❛➈❭➹❢➏❛➂❁➈❄➄❛➈✵➶❩➋♠➐✚➌✚➎❁➄☎➆✥➉❛➈✵➄❛➏❛➉➍➃❸➊✍➏✍➹✻➃❸➌⑧➏❛➂❩➆✥➏ ➏❛➂✚➈✍➘☎➊✥➆❫➌☎➶❩➈✹➉❛➈✥➆✍➒✇➃❸à✍➈✍➎❋➪✖➃❸➏❛➂❋➆✹➄❛➈✍➏✖➋✥➑✿➛❤➜➍➙r➙r➞●❑✖➄❛➋♠➐✚➉❛➊✍➈✍➄❳➪❳➂❼➃❸➊r➂✪✃❭➈✍➈✍➅ ➏❛➂✚➈➡➃❸➉✠➏❛➋❭✃♠➈✍➌✏➶❁➐✚➊➡✃❭➈✍➏❛➄✠➈❄➔✵➅❁➏❏➘♠➻ ➁➴➂✚➈✠➅✚➉❛➈✍➚♠➃❸➋❭➐❁➄❳➅➇➆❫➉r➆❫Ð❭➉r➆❫➅✚➂❁➄Ò➎❁➈❄➄❛➊✍➉➍➃❸➶❩➈✹➼✛✂ã➃❸➌☎➏❛➂❁➈❳➱❩➐❤➃❸➎✆➱➇➋✥➪ ➆❫➅✚➅✚➉❛➋❫➬♠➃❸➔➾➆❫➏➍➃❸➋♠➌☛➻ ✫✠➌✚➈➴➊★➆❫➌✵➎✚➈➡➽➇➌❁➈Ò➏❛➂❁➈✖➅➇➆❫➊r✃♠➈❄➏➍➃❸à✍➈✍➎✪➚❭➈✍➉❛➄➍➃❸➋♠➌☎➋✥➑☛➏❛➂❁➈ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔❅➃❸➌✞➆✯➄❛➏❛➉r➆★➃❸Ð❭➂❤➏➍➑❦➋♠➉❛➪❳➆✥➉❛➎Ñ➔➾➆❫➌✚➌❁➈✍➉✍➻Òá❳➈➡➽➇➌❁➈✸✷✺✹✖ ➭ ✕ ➺Ô➑❦➋❭➉Ò➆★➒✇➒ ✕ ✶ ✕ ✹✖ ➆✥➄✹➏❛➂❁➈✵➌❤➐✚➔✂➶❩➈✍➉✹➋❫➑Ô➶❼➃❸➏❛➄✹➏❛➂➇➆❫➏✹➂➇➆✍➚♠➈☎➶❩➈✍➈✍➌✞➄❛➈✍➉❛➚♠➃❸➊✍➈❄➎Ñ➑❦➉❛➋♠➔ ➏❛➂✚➈✠➱❩➋✥➪ ✻Ñ➶❩➈✍➏❏➪✖➈✍➈✍➌✏➏❛➂❁➈✹➏➍➃❸➔✵➈❄➄ ✕ ✹✖ ➆❫➌✚➎ ✕ ➻✯☎Ò➄❛➄❛➋❤➊➡➃❷➆❫➏❛➈✂➪✖➃❸➏❛➂❋➈✥➆❫➊r➂ ➅➇➆❫➊r✃♠➈✍➏Ò➏❛➂❁➈✠➑❦➐✚➌✚➊✍➏➍➃❸➋♠➌✺✹✖✹ ➭ ✕ ➺ ✥✛➭✟☛✹ ➭ ✕ ✹✖ ➺ ✲ ✷✖✹ ➭ ✕ ➺❛➺✝✆ ✠ ✹ ➪➴➂✚➈✍➉❛➈✂➪✖➈ ➏r➆❫✃❭➈☎➏❛➂❁➈✵➉➍➃❸Ð❭➂❤➏➍➓❏➂❩➆✥➌❁➎▼➒✇➃❸➔✯➃❸➏☎➋❫➑✻☛ ✰❢➏❛➂❤➃❸➄✯➌❤➐✚➔✂➶❩➈✍➉✹➃❸➄✯➏❛➂✚➈✵➒❸➈✍➚♠➈➡➒✦➋❫➑ ➶➇➆❫➊r✃♠➒❸➋♠Ð❋➆✥➂❁➈★➆❫➎❋➋❫➑✿➏❛➂✚➈✂➅❩➆✥➊➡✃❭➈✍➏✏✗✻➃❸➌❋➏❛➂❁➈✹➱➇➋❫➪ ✻ ✳ ➄✖➷❤➐❁➈❄➐❁➈✵➎❤➃❸➚♠➃❸➎✚➈✍➎ ➶❤➘☎➏❛➂❁➈✹➱➇➋❫➪✗✳ ➄❳➄❛➂❩➆✥➉❛➈✹➋❫➑✿➏❛➂✚➈✠➒✇➃❸➌❁✃❩➹❱➆❫➌✚➎❋➊✥➆❫➌✏➶❩➈✂➏❛➂❁➋❭➐❁Ð❭➂❤➏✠➋❫➑✦➆✥➄Ò➆❫➌ ➙✄✂➡Ï❩➙r➤✥↕❦➙➡➞✖➎✚➈➡➒❷➆✍➘✪➐✚➌❤➏➍➃✇➒❢➎✚➈✍➅➇➆❫➉❛➏❛➐✚➉❛➈P➑❦➋❭➉✖➏❛➂✚➈✠➒❷➆❫➄❛➏❳➶❼➃❸➏❳➃❸➌✪➏❛➂✚➈✹➅❩➆✥➊➡✃❭➈✍➏✍➻ ➁❳➂❁➈✞➅➇➆❫➊r✃♠➈✍➏➍➃❸à✍➈❄➎⑤➚❭➈✍➉❛➄➍➃❸➋♠➌✫➋❫➑✲➼✛✂ ➃❸➄✪➔✲➈✍➉❛➈➡➒❸➘♠➹➮➆❫➏ ➆❫➌❤➘✸➏➍➃❸➔✲➈ ✕ ➪❳➂❁➈❄➌✪➏❛➂✚➈✹➌❁➈✍➬❤➏Ò➅❩➆✥➊➡✃❭➈✍➏❳➏❛➋✵➶❩➈✹➏❛➉r➆✥➌❁➄❛➔✯➃❸➏❛➏❛➈❄➎❋➔✯➐✚➄❛➏❳➶❩➈✯➊r➂✚➋♠➄❛➈✍➌☛➹❩➏❛➋ ➄❛➈➡➒❸➈✍➊❄➏✵➏❛➂❁➈☎➅➇➆❫➊r✃♠➈✍➏✂➪✖➃❸➏❛➂▼➏❛➂✚➈✆➔✂➃❸➌❼➃❸➔☎➆✍➒☞✹✏✹✖ ➭ ✕ ➺➡➻✪➁❳➂❤➃❸➄✵➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➃❸➄✠➊★➆★➒✇➒❸➈✍➎⑧➆✯➏➍➃❸➔✵➈➡➓❏➄❛➏r➆✥➔✵➅❋➶❩➆✥➄❛➈✍➎❋➄❛➊r➂❁➈✍➔✲➈✹➶❩➈✍➊✥➆✥➐❁➄❛➈✂➏❛➂❁➈❄➉❛➈P➃❸➄Ò➆❫➌✞➆✍➒✇➓ ➏❛➈✍➉❛➌➇➆❫➏➍➃❸➚❭➈✹➶❁➐✚➏✳➈❄➷❤➐❼➃❸➚❭➆★➒❸➈❄➌❤➏➴➑❦➋❭➉❛➔✯➐❤➒❷➆❫➏➍➃❸➋❭➌✪➃❸➌✆➪❳➂❤➃❸➊➡➂☎➈✥➆✥➊➡➂☎➅➇➆❫➊r✃♠➈✍➏✿➃❸➄ ➄❛➏r➆❫➔✲➅❩➈✍➎❞➪✖➃❸➏❛➂✸➆▼↕❦➢✔✂➙✏▲➍➥❄↕❦➝✆✔ÔÏ ➆❫➄✹➃❸➏☎➆❫➉❛➉➍➃❸➚❭➈✍➄☎➆❫➌✚➎❞➏❛➂❁➈❄➌▼➅❩➆✥➊➡✃❭➈✍➏❛➄ ➆❫➉❛➈Ò➏❛➉r➆❫➌✚➄❛➔✯➃❸➏❛➏❛➈✍➎✪➃❸➌☎➃❸➌❁➊✍➉❛➈★➆❫➄➍➃❸➌✚Ð✪➋♠➉❛➎✚➈✍➉✖➋✥➑❢➏➍➃❸➔✵➈r➓➍➄❛➏r➆✥➔✵➅✚➄☛✰❩➄❛➈✍➈ ✬Ô➈r➑③➓ ➈✍➉❛➈✍➌✚➊✍➈✍➄✆✍✟❁➹ ✮✖✕✭✎✿➑❦➋❭➉✹➎❁➈✍➏r➆✍➃✇➒❸➄✵➋♠➌✞➏❛➂❤➃❸➄✯➑❦➋♠➉❛➔✂➐❤➒❷➆❫➏➍➃❸➋♠➌☛➻ ➵❢➆❫➉❛➈✍✃❤➂❞➆✥➌❁➎ ✑✹➆★➒✇➒❷➆✥Ð♠➈✍➉✹➂➇➆✍➚♠➈✹➄❛➂✚➋✥➪➴➌❋➏❛➂➇➆❫➏Ò➆✯➶➇➋♠➐✚➌❁➎☛➹☛➄➍➃❸➔✯➃✇➒❷➆❫➉✹➏❛➋✲➏❛➂❁➈✠➱➇➐❼➃❸➎✪➱➇➋✥➪ ➶❩➋❭➐❁➌✚➎✧➹Ô➆❫➅✚➅❼➒✇➃❸➈❄➄✵➏❛➋❋➏❛➂❼➃❸➄✲➅❩➆✥➊➡✃❭➈✍➏➍➃❸à✍➈✍➎→➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✒➆❫➄✂➪✖➈➡➒✇➒❦➻✞ß❳➋✥➪Ô➓ ➈✍➚❭➈✍➉✍➹☛➏❛➂❁➈✹➑❦➋❭➉❛➔✯➐❤➒❷➆❫➈✯➑s➋♠➉✠➏❛➂✚➈✂➎❁➈➡➒❷➆✍➘❤➄✠➃❸➌✞➏❛➂❁➈✂➅➇➆❫➊r✃♠➈✍➏➍➃❸à✍➈❄➎▼➊✥➆❫➄❛➈✲➆❫➉❛➈ ➄➍➃❸Ð♠➌❤➃✇➽➇➊✥➆❫➌❤➏➍➒❷➘✵➔✵➋❭➉❛➈✳➊✍➋❭➔✵➅❤➒✇➃❸➊✥➆❫➏❛➈✍➎✍✰❩➄❛➈❄➈ ✬✖➈➡➑❦➈❄➉❛➈✍➌❁➊❄➈ ✍ ✮✬✕✭✎❤➑❦➋❭➉✻➎✚➈✍➏r➆★➃✇➒❸➄✍➻ ➁❢➋✹➐❁➌✚➎❁➈❄➉❛➄❛➏r➆❫➌✚➎❋➏❛➂❁➈Ò➉❛➈➡➒❷➆❫➏➍➃❸➋❭➌❋➶❩➈✍➏➍➪✦➈✍➈✍➌☎➏❛➂❁➈Ò➊➡➒❸➋❤➊r✃✆➉r➆✥➏❛➈✏✠♠➹✚➏❛➂❁➈ ➶✚➐❁➊r✃♠➈✍➏❳➄➍➃❸à✍➈✸☛✍➭✒✠✥➺❳➆❫➌✚➎✪➏❛➂✚➈✠➉❛➈✍➄❛➐❤➒❸➏r➆❫➌❤➏❳➎✚➈➡➒❷➆✍➘♠➹☛➊✍➋♠➌✚➄➍➃❸➎✚➈✍➉❳➪➴➂➇➆❫➏❳➂➇➆❫➅❤➓ ➅❩➈✍➌✚➄✪➏❛➋⑧➆Ñ➶✚➐❁➉❛➄❛➏☎➋✥➑✂➅❩➆✥➊➡✃❭➈✍➏❛➄✍➻❑➁❳➂❁➈❋➅➇➆❫➊r✃♠➈❄➏✆➏❛➂➇➆❫➏☎➉❛➈✍➊❄➈➡➃❸➚♠➈✍➄✏➏❛➂❁➈ ➂❤➃❸Ð♠➂✚➈✍➄❛➏☎➷❤➐❁➈✍➐✚➈➡➃❸➌✚Ð❞➎❁➈➡➒❷➆✍➘❞➃❸➄✵➏❛➂❁➈☎➒❷➆❫➄❛➏☎➅➇➆❫➊r✃♠➈✍➏✵➋✥➑✂➆❋➶❁➐✚➉❛➄❛➏✍➻❞➁❳➂❁➈ ➶❩➋❭➐❁➌✚➎▼➋♠➌✞➏❛➂✚➈❃✲➍➃❸➏❛➏❛➈✍➉✠➃❸➄✹➅✚➉❛➋♠➅❩➋❭➉❛➏➍➃❸➋♠➌➇➆★➒Ò➏❛➋☎➏❛➂❁➈✲➄➍➃❸à✍➈✵➋❫➑Ô➏❛➂❁➈✵➶✚➐✚➉❛➄❛➏ ➆❫➌✚➎✪➃❸➌❤➚♠➈❄➉❛➄❛➈➡➒❸➘❋➅❁➉❛➋❭➅❩➋♠➉❛➏➍➃❸➋❭➌❩➆✍➒✖➏❛➋✲➏❛➂❁➈✂➊➡➒❸➋❤➊r✃✪➉r➆❫➏❛➈❭➻✳➁❳➂✚➈✹➔✵➈✥➆✥➌❁➄Ò➶❤➘ ➪❳➂❼➃❸➊r➂✸➏❛➂✚➈❋➄❛➋♠➐✚➉❛➊✍➈❋➊✥➆✥➌Ý➃❸➔✵➅✚➉❛➋✥➚♠➈❋➏❛➂✚➈❋➪✖➋♠➉❛➄❛➏✲➊✥➆❫➄❛➈✏➶❩➋♠➐✚➌❁➎✸➃❸➄☎➏❛➋ ➃❸➌✚➊✍➉❛➈✥➆❫➄❛➈✠➃❸➏❛➄❳➊➡➒❸➋❤➊r✃☎➉r➆❫➏❛➈ ✠✠➏❛➋✯➅❩➈❄➉❛➔✯➃❸➏❳➏❛➂❁➈Ò➶❁➐✚➉❛➄❛➏✖➏❛➋✂➅❩➆✥➄❛➄➴➏❛➂✚➉❛➋♠➐✚Ð♠➂ ➏❛➂✚➈✹➌❁➈✍➏❏➪✖➋♠➉❛✃☎➔✵➋❭➉❛➈✯➷❤➐❤➃❸➊r✃♠➒❸➘♠➻ ➲ä➃❸➌✚➊✍➈✞➏❛➂❁➈❋➶❩➋❭➐❁➌✚➎✚➄✪Ð❫➃❸➚❭➈✍➌✸➃❸➌❑Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❑➄❛➈✍➉❛➚♠➃❸➊❄➈Ñ➔✂➐❁➄❛➏☎➶❩➈ ➪✖➋❭➉❛➄❛➏➍➓➍➊★➆❫➄❛➈✠➶➇➋♠➐✚➌❁➎✚➄✹➭③➃❦➻ ➈♠➻✿➏❛➂✚➈❳➶❩➋♠➐✚➌❁➎✚➄✠➔✂➐✚➄❛➏✖➆✥➅❁➅❤➒❸➘✆➑❦➋❭➉✖➆✍➒✇➒✰➅❩➋♠➄❛➄➍➃✇➓ ➶❤➒❸➈✯➶❩➈❄➂❩➆✍➚♠➃❸➋♠➉❛➄✹➋✥➑✿➏❛➂❁➈✹➋❭➏❛➂❁➈❄➉➴➄❛➋❭➐❁➉❛➊❄➈✍➄r➺➡➹☛➏❛➂❁➈✹➅✚➉➍➃❸➔☎➆✥➉❛➘✆➑❦➐✚➌❁➊❄➏➍➃❸➋♠➌❋➋❫➑ ➆✪➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð❞➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➎✚➈✍➄➍➃❸Ð♠➌✚➈✍➎❞➏❛➋❋➎✚➈➡➒✇➃❸➚♠➈✍➉✲Ð♠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎❞➄❛➈✍➉➍➓ ➚♠➃❸➊✍➈✵➃❸➄✂➏❛➋❞➢ ➥✍➫✖✥❸➝❤↕❦➙✖➱❩➋✥➪❳➄✠➑❦➉❛➋❭➔ ➈★➆❫➊r➂▼➋♠➏❛➂✚➈✍➉✍➹❢➄❛➋✏➏❛➂❩➆✥➏✵➆✵➱❩➋✥➪✔➊★➆❫➌ ➂➇➆★➚❭➈❳➋♠➌❤➒❸➘✵➆✖➒✇➃❸➔✯➃❸➏❛➈✍➎☎➌❁➈❄Ð❤➆❫➏➍➃❸➚❭➈✠➈➡Þ❙➈✍➊✍➏❢➋♠➌✲➋♠➏❛➂✚➈✍➉✻➱❩➋✥➪❳➄✍➻❢➁❳➂✚➈✖➼✛✂ ➄❛➊r➂❁➈✍➔✲➈Ô➃❸➄❛➋✥➒❷➆❫➏❛➈✍➄✦➈✥➆❫➊r➂✯➄❛➋❭➐❁➉❛➊❄➈Ô➑s➉❛➋♠➔✛➏❛➂✚➈✖➋♠➏❛➂✚➈✍➉❛➄✿➶❤➘✯➅✚➉❛➋✥➚♠➃❸➎❼➃❸➌✚Ð✯➃❸➏✿➆ ➄❛➅❩➈✍➊r➃✇➽❩➈✍➎❋➄❛➂➇➆❫➉❛➈✹➋✥➑❢➏❛➂✚➈✹➶❩➆✥➌❁➎✚➪✖➃❸➎❁➏❛➂❋➐✚➌✚➎❁➈✍➉Ò➋✥➚♠➈✍➉➍➒❸➋ä➆❫➎❋➊✍➋❭➌❁➎❤➃❸➏➍➃❸➋♠➌✚➄✍➻ ➁❳➂❁➈❋➪✦➋♠➉❛✃▼➋✥➑✲➵❢➆❫➉❛➈✍✃❤➂ ➆❫➌✚➎ ✑✹➆★➒✇➒❷➆✥Ð♠➈✍➉❋➅✚➉❛➋✥➚♠➃❸➎❁➈❄➄❋➆Ñ➪❳➆✍➘▼➑❦➋❭➉✵➏❛➂❁➈ ➄❛➋♠➐✚➉❛➊✍➈☎➏❛➋✪➊❄➋♠➔✵➅✚➐❁➏❛➈☎➏❛➂✚➈☎➔☎➆❫➬♠➃❸➔✂➐❁➔❐➷❤➐✚➈✍➐✚➈➡➃❸➌❁Ð⑧➎❁➈➡➒❷➆✍➘✞➪❳➂❼➃❸➊r➂▼➃❸➏❛➄ ➅➇➆❫➊r✃♠➈✍➏❛➄✖➪✖➃✇➒✇➒❱➈✍➌❁➊❄➋♠➐✚➌❤➏❛➈✍➉✍➹✚➅❁➉❛➋✥➚♠➃❸➎✚➈✍➎❋➏❛➂❩➆✥➏✖➏❛➂✚➈✠➄❛➋♠➐✚➉❛➊✍➈Ò➉❛➈✍➄❛➏❛➉➍➃❸➊✍➏❛➄✖➃❸➏➍➓ ➄❛➈➡➒✇➑✰➏❛➋✹➆✹➭✒✠✭✓ ☛➡➺✻➏❛➋❭✃♠➈✍➌✵➶✚➐✚➊➡✃❭➈✍➏ ➽✚➒❸➏❛➈✍➉✍➻ ✎✳➐✚➏❢➏❛➂❁➈✦➌❁➈✍➏❏➪✖➋♠➉❛✃ ✳ ➄✿➄❛➊➡➂✚➈✍➎✚➐❼➒✇➓ ➃❸➌✚Ð☎➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➎✚➋❤➈✍➄✖➌✚➋♠➏✖➎✚➈✍➅❩➈❄➌❁➎❋➋❭➌✆➏❛➂❤➃❸➄✖➽❁➒❸➏❛➈❄➉✍➻✳➯❏➌❁➎✚➈✍➈✍➎☛➹☛➆✥➌✵➃❸➔✯➓ ➅❩➋❭➉❛➏r➆❫➌❤➏✹➅❩➋✥➃❸➌❤➏✹➆❫➶➇➋♠➐✚➏✹➏❛➂❼➃❸➄✠➑❦➋❭➉❛➔✾➋❫➑✦Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈✯➃❸➄Ò➏❛➂❩➆✥➏ ➏❛➂✚➈✹➏❛➉r➆★Ó☎➊✠➽✚➒❸➏❛➈✍➉❛➄✠➎✚➋☎➌❁➋❭➏❳➅❼➒❷➆✍➘❋➆✥➌❤➘☎➉❛➋❫➒❸➈✹➃❸➌❋➅❩➆✥➊➡✃❭➈✍➏✠➄❛➊r➂❁➈❄➎❁➐❤➒✇➃❸➌❁Ð✚➻ ✑➴➃❸➚♠➈✍➌❞➏❛➂➇➆❫➏✂➏❛➂❁➈☎➼✛✂ ➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➊✥➆✥➌▼➎✚➈➡➒✇➃❸➚♠➈✍➉☎Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈✍➎ ➄❛➈✍➉❛➚♠➃❸➊✍➈❭➹❁➪❳➂❤➘✂➌❁➋❭➏✖➐✚➄❛➈➴➏❛➂❤➃❸➄❳➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✾➏❛➋✹➅✚➉❛➋❫➚♠➃❸➎✚➈✹➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎☎➄❛➈✍➉➍➓ ➚♠➃❸➊✍➈✲➆❫➄❳➪✖➈➡➒✇➒ ✬✖➈✍➊✥➆✍➒✇➒❢➏❛➂➇➆❫➏❳➏❛➂❁➈✹Ð❭➋❤➆✍➒❢➋❫➑❱Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❋➄❛➈✍➉❛➚♠➃❸➊✍➈✂➪❳➆✥➄ ➏❛➋✵➅✚➉❛➋✥➚♠➃❸➎❁➈✵➪✦➋♠➉❛➄❛➏➍➓❏➊✥➆❫➄❛➈✂Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➄✍➹☛➪❳➂❤➃❸➊➡➂✞➆✥➉❛➈✯➉❛➈✍➷❤➐❤➃❸➉❛➈✍➎✞➶❤➘✆➃❸➌❤➓ ➏❛➋❫➒❸➈❄➉r➆❫➌❤➏✦➆❫➌✚➎✯➉➍➃❸Ð✥➃❸➎☎➆❫➅✚➅❼➒✇➃❸➊★➆❫➏➍➃❸➋♠➌➇➄✍➻✖➁❳➂❁➈✦Ð♠➋❤➆✍➒❩➋✥➑❩➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✵➄❛➈❄➉❛➚♠➃❸➊✍➈ ➃❸➄❢➌✚➋♠➏❢➏❛➋✠➅✚➉❛➋✥➚♠➃❸➎❁➈❳➪✖➋❭➉❛➄❛➏➍➓❏➊✥➆❫➄❛➈✖➎✚➈➡➒❷➆✍➘✹➶❩➋❭➐❁➌✚➎❁➄✍➹✚➶✚➐❁➏❢➉r➆✥➏❛➂❁➈❄➉❢➏❛➋✠➔✯➃❸➌❤➓ ➃❸➔✯➃❸à✍➈❋➏❛➂✚➈❋➆✥➊✍➏❛➐➇➆★➒❳Ï➇➫♠➥✍↕ ✟✍➝❤➤✥↕◗➫✵➎✚➈➡➒❷➆✍➘✸➶➇➋♠➐✚➌❁➎✚➄☎➏❛➂❁➈✍➉❛➈❄➶❤➘✸➆✍➒✇➒❸➋✥➪Ô➃❸➌✚Ð ➏❛➋❫➒❸➈❄➉r➆❫➌❤➏Ò➆❫➌✚➎❋➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈✵➆✥➅❁➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❁➄✹➏❛➋✂➉❛➈✍➈➡➒☛➃❸➌☎➏❛➂✚➈➡➃❸➉❳➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃ ➅❩➋✥➃❸➌❤➏✍➻❑➯➍➌→➏❛➂❁➈✏➌❁➈✍➬❤➏☎➄❛➈✍➊❄➏➍➃❸➋♠➌☛➹❳➪✖➈✪➄❛➂✚➋✥➪ ➏❛➂➇➆❫➏✵➏❛➂✚➈❋➼✛✂❅➆★➒❸Ð❭➋❫➓
rithm, with its em phasis on isolation, is not well suited for O prov C Scheduling Als orithms gor Predicted Service Table Q The mean and th percentile queueing de lay We motivate the development of our scheduling algorithm (m t conside a single-link network carrying some numbe assume that all sources conform to some traffic filter such as the token buckets described above furt hermore, assume that all the clients are bursty se and wish to mix their affic so that in the aggregate achieve a better u be moved out as far to accommodate the jitter induced of bandwidth and a cont rolled delay. How does one best by the burst. Furthermore, the particular source produc- ing the burst is not singled out for increased jitter; all the schedule the packets to achieve low post acto delay bounds sources share in all the jitter induced by the bursts of all the (or, equivalently, minimal play-back points What behavior does the wF. algorithm induce? whe sources. ecall that when the packets are of uniform size ever there is a backlog in the queue, packets leave the queue the total queueing delay in any time period (summed over t rates proportional to their clock rates. Consider a mo- inde pendent of the sched uling algorit hm. The ment when all sources are transmit ting uniformly at their FIFO algorithm splits this delay evenly, whereas the W clock rates except for one which emits a burst of packets lgorit hm assigns the delay to the fows that caused the mo The Wf- algorithm would continue to send the packet mentary queueing (by sending bursts). When the delays ar from the uniform sources at their clock rates, so their pack shared as in FIFO, in what might be called a multiplexing ts are not queued for any significant time whereas the back- of bursts, the post tacto jitter bounds are smaller than when log of packets from the bursty source will take a long tim the sources are isolated from each other as in wF., This to drain. Thus, a burst by one source causes a sharp in was exactly our goal; under the same link utilization, FIFO crease in the delay seen by that source, and has minimal allows a number of sources aggregating their traf effects on the delays seen by the ot her sources. The mean a lower overall delay jitter. delay will be rather low, assuming the net work link is not In order to test our intuition we have simulated bot h ver-committed, but a burst will induce jit ter directly, and the WF- and FIFO algorit hms. The g ppendix contains a mostly, affecting only the source that emit ted the bur complete description of our simulation procedure, we only WF- provides for a great degree of isolat present the results here. We consider a single Q mbit gsec sources are protected from other sources' bursts. Is this the link being utilized by ct flows, each having the same statis best approach to obtaining the lowest play-back point when tical generation process. In Table Qwe show the mean and th percentile queueing delays for a sample fow (the a number of sources are sharing a link? We argue that this dat a from the various flows are similar) under each of the isolation, while necessary for providing guaranteed service nterproduct ive for predic ted service t wo schedul ithms. Note that while the The nature of play-back real-time applications allows the lays are a bout the same for the two algorit hms, the scheduling algorithm to delay all packets up to the play-back percentile delays are significantly smaller under the FIFO point without adversely affecting the application's perfor algorit hm. This confirms our anal The FIFo queue discipline has generally been con deadline. For such problems, the standard earliest -deadline- ered ineffective for providing, real-time service, in fact, it first scheduling algorithm, as described in eference [Q B has been proven optimal. However, in our gedanken erperiment possible algorithm ([QB. The reason is that if ne source he play-back points are not set a priori, as in the above cts excessive traffic into the net, this disrupts the service for everyone. This assessment, howe ver, arises from a fail eference, but are rather the result of the clients adapting to the current level of delay. to dist inguish the two separate objectives of any traffic con Let us consider a simple example where a class of clent trol algorithm, iso= tion and sharing. Isolation is the more have similar service desires. This that they are all fundamental goal; it provides guaranteed service for well t isfied with the they all have the behaved clients and quarantines misbehaving sources. ]ut same play-back point and thus the deadline. if the sharing if it is performed in the context of an encom passing deadline for each packet is a constant offset to the arrival duling algorithm b traffic from different IFO; the packet that is closest to its deadline is the all; bursts are multiplexed so that the post facto jitter is ne that arrived first. Hyman, Lazar, and pacifici, in ef aller for everyone. The FIFo scheme is an effective shar- rence [gyB also make this obse that FiFo is ing scheme, but it d ot provide olation w the ot her hand, is an effective method for isolat ion If e the fifo anize the traffic into classes of clients with similar service discipline instead of wF requirements, we find that this reasoning leads t rives, this burst through th scheme in which the queuing decision is in two steps: a first hile subsequent packets from the other source step to insure isolation of cla smaller tb rarily delayed; this latter delay, howe ver, is much thod wit hin n the delay that the bursting source wo uld have FIFO is not the only interesting sharing method. gn WF-. Thus, the play-back poi ot her sharing me thod is priority, which has a very different behavior than FIFO. The goal of FIFo is to let every source
➉➍➃❸➏❛➂✚➔✪➹✰➪✖➃❸➏❛➂✪➃❸➏❛➄✹➈✍➔✵➅✚➂❩➆✥➄➍➃❸➄✯➋❭➌✪➃❸➄❛➋❫➒❷➆❫➏➍➃❸➋❭➌✧➹❢➃❸➄✹➌✚➋♠➏Ò➪✖➈➡➒✇➒✿➄❛➐❤➃❸➏❛➈✍➎ ➑s➋♠➉ ➏❛➂❤➃❸➄✹➏r➆❫➄❛✃❩➻✦➁❳➂❤➐❁➄✍➹✰➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✵➄✂➄❛➅❩➈✍➊➡➃✇➽➇➊✥➆✍➒✇➒❸➘▼➎❁➈❄➄➍➃❸Ð♠➌✚➈✍➎Ñ➏❛➋☎➅✚➉❛➋✥➚♠➃❸➎❁➈ ➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈✂➔✯➐✚➄❛➏❳➶❩➈✠➑❦➋❭➐❁➌✚➎✧➻ ✏❙➀✬❚●✺♠Ú➇Û✻❂ Ü❸Ø✆☎✏❹❏❂ ☎✚Ù✚❾◗Ü✇❽ ❚❃❆✯❻❏✽➍Ù✚❾ ✷✻❾✺♠Ú✚Ü✇➀❫❽✢✺❭Ú☞✏❉✺❤❾ ✒❭Ü✇➀✔✺ ➼✞➈☎➔✵➋♠➏➍➃❸➚❭➆✥➏❛➈✪➏❛➂✚➈✆➎✚➈✍➚❭➈➡➒❸➋♠➅✚➔✵➈✍➌❤➏☎➋❫➑Ò➋❭➐❁➉✂➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð→➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➶❤➘✲➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉➍➃❸➌❁Ð☎➏❛➂✚➈➴➑❦➋✥➒✇➒❸➋❫➪✖➃❸➌✚Ð✏➛❼➙r➞❤➝❤↔❤➧❤➙✥↔✯➙✄✂➡Ï➇➙✥➜➡➢✔✂➙✥↔❩↕③➻ ✖➋♠➌✚➄➍➃❸➎❁➈✍➉ ➆☎➄➍➃❸➌❁Ð✥➒❸➈➡➓❦➒✇➃❸➌❁✃❞➌✚➈✍➏➍➪✦➋♠➉❛✃❋➊✥➆✥➉❛➉❛➘♠➃❸➌❁Ð❋➄❛➋❭➔✵➈✵➌❤➐✚➔✯➶❩➈❄➉✯➋✥➑✖➊➡➒✇➃❸➈❄➌❤➏❛➄✍➹✖➆✥➌❁➎ ➆✥➄❛➄❛➐❁➔✲➈✪➏❛➂➇➆❫➏✲➆★➒✇➒❳➄❛➋❭➐❁➉❛➊✍➈❄➄✵➊✍➋♠➌❤➑❦➋❭➉❛➔ ➏❛➋❋➄❛➋❭➔✵➈☎➏❛➉r➆★Ó☎➊✵➽✚➒❸➏❛➈✍➉✵➄❛➐✚➊➡➂ ➆✥➄✖➏❛➂✚➈✠➏❛➋♠✃❭➈✍➌✪➶✚➐✚➊➡✃❭➈✍➏❛➄❳➎❁➈✍➄❛➊❄➉➍➃❸➶❩➈✍➎✞➆❫➶➇➋❫➚❭➈♠➻★✛❁➐✚➉❛➏❛➂❁➈✍➉❛➔✲➋♠➉❛➈♠➹❙➆❫➄❛➄❛➐✚➔✵➈ ➏❛➂➇➆❫➏Ô➆★➒✇➒✰➏❛➂❁➈❳➊➡➒✇➃❸➈❄➌❤➏❛➄✠➆✥➉❛➈✠➶✚➐❁➉❛➄❛➏❏➘✵➄❛➋♠➐✚➉❛➊✍➈✍➄✍➹➇➆✥➌❁➎✏➪Ô➃❸➄❛➂☎➏❛➋✹➔✯➃❸➬✲➏❛➂❁➈➡➃❸➉ ➏❛➉r➆✍Ó☎➊❋➄❛➋❞➏❛➂❩➆✥➏✵➃❸➌❑➏❛➂❁➈✞➆✥Ð♠Ð♠➉❛➈❄Ð❤➆❫➏❛➈✞➏❛➂✚➈✍➘❑➆✥➊➡➂❤➃❸➈✍➚♠➈⑧➆Ñ➶❩➈❄➏❛➏❛➈✍➉✪➐✚➄❛➈ ➋✥➑✹➶➇➆❫➌✚➎❁➪✖➃❸➎✚➏❛➂✛➆✥➌❁➎ ➆Ñ➊✍➋❭➌❤➏❛➉❛➋❫➒✇➒❸➈✍➎❑➎✚➈➡➒❷➆✍➘♠➻✷ß➴➋✥➪✾➎✚➋❤➈✍➄☎➋♠➌✚➈❋➶❩➈✍➄❛➏ ➄❛➊r➂✚➈✍➎❁➐❤➒❸➈✵➏❛➂✚➈✹➅❩➆✥➊➡✃❭➈✍➏❛➄❳➏❛➋☎➆❫➊r➂❤➃❸➈✍➚♠➈✹➒❸➋✥➪ãÏ➇➫♠➥❄↕✆✟✍➝❤➤✥↕◗➫❳➎✚➈➡➒❷➆✍➘✪➶❩➋❭➐❁➌✚➎❁➄ ➭❦➋❭➉✍➹❩➈✍➷❤➐❤➃❸➚❭➆★➒❸➈✍➌❤➏➍➒❸➘❭➹❢➔✯➃❸➌❤➃❸➔☎➆★➒✿➅❤➒❷➆✍➘♠➓➍➶➇➆❫➊r✃❋➅❩➋✥➃❸➌❤➏❛➄r➺✁ ➼✸➂➇➆❫➏✦➶❩➈✍➂➇➆★➚♠➃❸➋❭➉✖➎✚➋❤➈✍➄✖➏❛➂✚➈➴➼✛✂❑➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✔➃❸➌✚➎❁➐✚➊✍➈✺❋➼✸➂✚➈✍➌❤➓ ➈❄➚♠➈✍➉✖➏❛➂✚➈✍➉❛➈✖➃❸➄❳➆✹➶❩➆✥➊➡✃♠➒❸➋❭Ð✯➃❸➌☎➏❛➂❁➈❳➷❤➐✚➈✍➐✚➈♠➹✚➅❩➆✥➊➡✃❭➈✍➏❛➄✖➒❸➈✥➆✍➚♠➈✠➏❛➂✚➈❳➷❤➐❁➈❄➐❁➈ ➆✥➏✯➉r➆✥➏❛➈✍➄✂➅❁➉❛➋❭➅❩➋♠➉❛➏➍➃❸➋❭➌❩➆✍➒✹➏❛➋❋➏❛➂✚➈➡➃❸➉✵➊➡➒❸➋❤➊r✃▼➉r➆❫➏❛➈❄➄✍➻✞✖➋♠➌✚➄➍➃❸➎❁➈✍➉➾➆✪➔✵➋❫➓ ➔✲➈✍➌❤➏✯➪❳➂✚➈✍➌→➆★➒✇➒❳➄❛➋❭➐❁➉❛➊❄➈✍➄☎➆❫➉❛➈✏➏❛➉r➆❫➌✚➄❛➔✯➃❸➏❛➏➍➃❸➌✚Ð▼➐❁➌❤➃✇➑❦➋❭➉❛➔✯➒❸➘→➆❫➏✂➏❛➂❁➈➡➃❸➉ ➊r➒❸➋❤➊➡✃❞➉r➆❫➏❛➈❄➄✯➈✍➬❤➊❄➈✍➅❁➏✂➑❦➋♠➉✂➋♠➌✚➈✪➪❳➂❤➃❸➊➡➂❞➈✍➔✯➃❸➏❛➄➾➆✪➶❁➐✚➉❛➄❛➏✵➋❫➑Ò➅➇➆❫➊r✃♠➈✍➏❛➄✍➻ ➁❳➂✚➈Ñ➼✛✂ ➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔❇➪✖➋❭➐❼➒❸➎✫➊✍➋❭➌❤➏➍➃❸➌❤➐❁➈▼➏❛➋❞➄❛➈✍➌✚➎✫➏❛➂❁➈✞➅❩➆✥➊➡✃❭➈✍➏❛➄ ➑s➉❛➋♠➔✔➏❛➂✚➈✹➐✚➌❼➃✇➑❦➋❭➉❛➔✾➄❛➋♠➐✚➉❛➊✍➈✍➄Ò➆❫➏✦➏❛➂❁➈➡➃❸➉Ò➊➡➒❸➋❤➊r✃✆➉r➆✥➏❛➈✍➄✍➹✚➄❛➋✵➏❛➂✚➈➡➃❸➉❳➅❩➆✥➊➡✃♠➓ ➈❄➏❛➄✿➆❫➉❛➈✖➌✚➋♠➏✿➷❤➐❁➈❄➐❁➈✍➎✂➑❦➋♠➉✿➆✥➌❤➘✹➄➍➃❸Ð♠➌❤➃✇➽❩➊✥➆✥➌❤➏❳➏➍➃❸➔✵➈✖➪❳➂❁➈❄➉❛➈✥➆❫➄✿➏❛➂❁➈✦➶❩➆✥➊➡✃♠➓ ➒❸➋❭Ð✪➋✥➑✖➅➇➆❫➊r✃♠➈✍➏❛➄✹➑❦➉❛➋❭➔ ➏❛➂✚➈✯➶✚➐✚➉❛➄❛➏➍➘❋➄❛➋♠➐✚➉❛➊✍➈✵➪✖➃✇➒✇➒✦➏r➆❫✃❭➈☎➆✯➒❸➋♠➌✚Ð❋➏➍➃❸➔✵➈ ➏❛➋✞➎✚➉r➆★➃❸➌☛➻✸➁❳➂❤➐✚➄✍➹❳➆❋➶❁➐✚➉❛➄❛➏☎➶❤➘▼➋♠➌✚➈✪➄❛➋❭➐❁➉❛➊✍➈❋➊★➆❫➐✚➄❛➈✍➄✏➆ ➄❛➂➇➆❫➉❛➅❞➃❸➌❼➓ ➊❄➉❛➈✥➆❫➄❛➈✪➃❸➌→➏❛➂❁➈✏➎❁➈➡➒❷➆✍➘→➄❛➈✍➈✍➌✸➶❤➘⑧➏❛➂❩➆✥➏✆➄❛➋❭➐❁➉❛➊❄➈♠➹❳➆✥➌❁➎→➂❩➆✥➄✆➔✂➃❸➌❼➃❸➔☎➆✍➒ ➈rÞ☛➈✍➊✍➏❛➄✂➋♠➌✞➏❛➂❁➈✲➎❁➈➡➒❷➆✍➘❤➄✵➄❛➈❄➈✍➌▼➶❤➘❋➏❛➂✚➈✵➋♠➏❛➂✚➈✍➉✂➄❛➋♠➐✚➉❛➊✍➈✍➄✍➻☎➁❳➂❁➈✲➔✵➈✥➆❫➌ ➎✚➈➡➒❷➆✍➘▼➪✖➃✇➒✇➒❳➶❩➈☎➉r➆✥➏❛➂❁➈✍➉✹➒❸➋✥➪✹➹✿➆❫➄❛➄❛➐✚➔✯➃❸➌✚Ð▼➏❛➂❁➈☎➌✚➈✍➏➍➪✦➋♠➉❛✃✪➒✇➃❸➌✚✃①➃❸➄✂➌✚➋♠➏ ➋✥➚❭➈✍➉➍➓❏➊✍➋♠➔✵➔✂➃❸➏❛➏❛➈✍➎✧➹✰➶✚➐❁➏✂➆✯➶✚➐✚➉❛➄❛➏✹➪Ô➃✇➒✇➒❢➃❸➌✚➎✚➐❁➊✍➈✗✲➍➃❸➏❛➏❛➈❄➉✠➎❤➃❸➉❛➈✍➊✍➏➍➒❸➘♠➹✿➆✥➌❁➎ ➔✲➋♠➄❛➏➍➒❸➘♠➹✰➆✍Þ☛➈✍➊✍➏➍➃❸➌✚Ð☎➋♠➌❤➒❸➘❋➏❛➂❁➈✹➄❛➋❭➐❁➉❛➊✍➈✹➏❛➂➇➆❫➏Ò➈✍➔✯➃❸➏❛➏❛➈✍➎✏➏❛➂❁➈✂➶✚➐❁➉❛➄❛➏✍➻ ➼✛✂ ➅✚➉❛➋✥➚♠➃❸➎❁➈✍➄☎➑❦➋♠➉✏➆▼Ð♠➉❛➈★➆❫➏☎➎❁➈❄Ð♠➉❛➈✍➈✞➋❫➑✹➃❸➄❛➋✥➒❷➆❫➏➍➃❸➋♠➌☛➹✂➄❛➋▼➏❛➂➇➆❫➏ ➄❛➋❭➐❁➉❛➊❄➈✍➄Ò➆❫➉❛➈✠➅✚➉❛➋♠➏❛➈❄➊✍➏❛➈✍➎✆➑s➉❛➋♠➔✔➋❭➏❛➂❁➈❄➉➴➄❛➋❭➐❁➉❛➊❄➈✍➄☛✳❩➶✚➐❁➉❛➄❛➏❛➄✍➻✿➯❏➄✖➏❛➂❼➃❸➄❳➏❛➂✚➈ ➶➇➈✍➄❛➏❳➆❫➅✚➅✚➉❛➋❤➆❫➊r➂✪➏❛➋✂➋♠➶✚➏r➆★➃❸➌❤➃❸➌✚Ð✆➏❛➂✚➈➴➒❸➋✥➪✖➈❄➄❛➏✖➅❼➒❷➆✍➘♠➓❏➶➇➆❫➊r✃✆➅➇➋❫➃❸➌❤➏❳➪❳➂❁➈✍➌ ➆✂➌❤➐❁➔✂➶❩➈✍➉Ò➋❫➑✿➄❛➋♠➐✚➉❛➊✍➈✍➄➮➆❫➉❛➈✂➄❛➂❩➆✥➉➍➃❸➌❁Ð ➆P➒✇➃❸➌✚✃✬①➼✞➈✂➆❫➉❛Ð❭➐❁➈✂➏❛➂➇➆❫➏❳➏❛➂❼➃❸➄ ➃❸➄❛➋✥➒❷➆❫➏➍➃❸➋❭➌✧➹✦➪➴➂❤➃✇➒❸➈☎➌❁➈❄➊✍➈✍➄❛➄r➆❫➉❛➘❋➑❦➋❭➉✹➅❁➉❛➋✥➚♠➃❸➎❤➃❸➌❁Ð✞Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎▼➄❛➈✍➉❛➚♠➃❸➊❄➈♠➹ ➃❸➄Ò➊✍➋♠➐✚➌❤➏❛➈✍➉❛➅✚➉❛➋❤➎❁➐✚➊✍➏➍➃❸➚❭➈➾➑s➋♠➉✖➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻ ➁❳➂❁➈✦➌❩➆✥➏❛➐❁➉❛➈❳➋❫➑❙➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃✲➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈Ò➆❫➅✚➅❤➒✇➃❸➊✥➆❫➏➍➃❸➋❭➌❩➄Ò➆★➒✇➒❸➋✥➪❳➄✳➏❛➂✚➈ ➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð☎➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➏❛➋Ò➎❁➈r➒❷➆★➘✲➆★➒✇➒➇➅➇➆❫➊r✃♠➈✍➏❛➄❢➐✚➅✯➏❛➋Ò➏❛➂❁➈✿➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃ ➅➇➋❫➃❸➌❤➏✪➪✖➃❸➏❛➂✚➋♠➐✚➏✏➆❫➎✚➚♠➈✍➉❛➄❛➈➡➒❸➘✫➆✍Þ☛➈✍➊✍➏➍➃❸➌✚Ð❞➏❛➂❁➈ ➆❫➅✚➅❼➒✇➃❸➊★➆❫➏➍➃❷➋♠➌✍✳ ➄❋➅❩➈❄➉➍➑❦➋♠➉➍➓ ➔➾➆❫➌✚➊✍➈♠➻✛➁❳➂❤➐✚➄✍➹❳➋♠➌✚➈Ñ➊★➆❫➌✸➏❛➂❤➃❸➌✚✃❑➋❫➑✹➏❛➂✚➈✞➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃❑➅➇➋❫➃❸➌❤➏❋➆✥➄✏➆ ➎✚➈✥➆✥➎❼➒✇➃❸➌✚➈♠➻ ✛❁➋❭➉✳➄❛➐✚➊➡➂✲➅❁➉❛➋❭➶❼➒❸➈✍➔✲➄✍➹❁➏❛➂✚➈❳➄❛➏r➆❫➌✚➎❩➆✥➉❛➎☎➈✥➆❫➉➍➒✇➃❸➈✍➄❛➏➍➓❏➎✚➈✥➆❫➎❤➒❸➃❸➌❩➈➡➓ ➽➇➉❛➄❛➏✻➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð✵➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✪➹➇➆❫➄✰➎❁➈❄➄❛➊✍➉➍➃❸➶❩➈✍➎✯➃❸➌❉✬✖➈➡➑s➈✍➉❛➈✍➌✚➊✍➈ ✍✒✂ ✣✎❏➹✥➂➇➆❫➄ ➶➇➈✍➈✍➌❋➅✚➉❛➋❫➚❭➈✍➌❋➋♠➅✚➏➍➃❸➔☎➆✍➒❦➻❳ß➴➋✥➪✖➈❄➚♠➈✍➉✍➹✚➃❸➌❋➋♠➐✚➉✠➛❤➙r➞❤➝❤↔❤➧❤➙✥↔✆➙✄✂➡Ï➇➙✥➜➡➢✔✂➙✥↔❩↕ ➏❛➂✚➈✪➅❤➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃→➅❩➋❫➃❸➌❤➏❛➄✏➆❫➉❛➈✏➌❁➋❭➏✵➄❛➈✍➏❋➝☎Ï✰➜➡➢③➫❤➜r➢ ➹✦➆❫➄✂➃❸➌Ý➏❛➂✚➈❋➆✥➶❩➋✥➚♠➈ ➉❛➈r➑❦➈✍➉❛➈✍➌✚➊✍➈♠➹✿➶✚➐❁➏✲➆❫➉❛➈☎➉r➆❫➏❛➂✚➈✍➉✂➏❛➂❁➈☎➉❛➈✍➄❛➐❤➒❸➏✂➋❫➑Ò➏❛➂❁➈☎➊➡➒✇➃❸➈❄➌❤➏❛➄☎➆❫➎➇➆❫➅✚➏➍➃❸➌❁Ð ➏❛➋✲➏❛➂❁➈✹➊✍➐✚➉❛➉❛➈✍➌❤➏❳➒❸➈✍➚♠➈➡➒❱➋❫➑✿➎✚➈➡➒❷➆✍➘♠➻ ✳✧➈❄➏✖➐❁➄✦➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉✠➆✹➄➍➃❸➔✵➅❤➒❸➈✹➈✍➬✚➆❫➔✲➅❼➒❸➈✂➪❳➂❁➈✍➉❛➈➮➆✠➊➡➒❷➆❫➄❛➄❳➋✥➑❢➊➡➒✇➃❸➈✍➌❤➏❛➄ ➂➇➆✍➚♠➈☎➄➍➃❸➔✯➃✇➒❷➆✥➉✵➄❛➈✍➉❛➚♠➃❸➊✍➈✏➎❁➈✍➄➍➃❸➉❛➈❄➄✍➻✪➁❳➂❤➃❸➄P➃❸➔✵➅❤➒✇➃❸➈✍➄✵➏❛➂➇➆❫➏✂➏❛➂✚➈✍➘▼➆✥➉❛➈☎➆★➒✇➒ ➄r➆✥➏➍➃❸➄➍➽❩➈❄➎Ý➪✖➃❸➏❛➂▼➏❛➂✚➈✵➄r➆✥➔✵➈☎➎❁➈➡➒❷➆✍➘ ✲❏➃❸➏❛➏❛➈✍➉☛✰❢➏❛➂❤➐✚➄✯➏❛➂✚➈✍➘⑧➆★➒✇➒✦➂❩➆✍➚♠➈☎➏❛➂✚➈ ➄r➆✥➔✵➈❋➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃✸➅❩➋✥➃❸➌❤➏✏➆❫➌✚➎✸➏❛➂❤➐✚➄✆➏❛➂✚➈❋➄r➆❫➔✲➈✪➎✚➈✥➆❫➎❤➒✇➃❸➌❁➈❭➻⑤➯❦➑✹➏❛➂✚➈ ➎✚➈✥➆✥➎❼➒✇➃❸➌✚➈ ➑s➋♠➉✂➈✥➆❫➊r➂▼➅➇➆❫➊r✃♠➈✍➏✯➃❸➄✲➆❋➊✍➋♠➌✚➄❛➏r➆❫➌❤➏✵➋✥Þ☛➄❛➈✍➏✂➏❛➋❋➏❛➂❁➈ ➆❫➉❛➉➍➃❸➚❭➆★➒ ➏➍➃❸➔✲➈♠➹❳➏❛➂❁➈✏➎❁➈★➆❫➎❤➒✇➃❸➌❁➈▼➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð✫➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✒➶❩➈✍➊❄➋♠➔✵➈✍➄✍➹❳➄❛➐✚➉❛➅✚➉➍➃❸➄➍➓ ➃❸➌✚Ð❫➒❸➘❭➹✯✛✻➯■✛✽✫✿✰✚➏❛➂❁➈✂➅➇➆❫➊r✃♠➈✍➏Ò➏❛➂❩➆✥➏Ô➃❸➄Ò➊➡➒❸➋♠➄❛➈❄➄❛➏✠➏❛➋✵➃❸➏❛➄❳➎❁➈✥➆✥➎❼➒✇➃❸➌✚➈✵➃❸➄✠➏❛➂✚➈ ➋❭➌❁➈✲➏❛➂❩➆✥➏✂➆❫➉❛➉➍➃❸➚♠➈❄➎①➽➇➉❛➄❛➏✍➻✲ß➴➘❤➔☎➆✥➌✧➹ ✳✻➆✥à✥➆❫➉✍➹ ➆❫➌✚➎▼➵❢➆✥➊➡➃✇➽❩➊r➃❦➹ ➃❸➌ ✬Ô➈r➑③➓ ➈❄➉❛➈✍➌❁➊❄➈ ✍✒✂✔✝✏✎➍➹➇➆★➒❸➄❛➋☎➔☎➆✥✃♠➈✹➏❛➂❤➃❸➄✠➋❭➶❁➄❛➈✍➉❛➚❭➆✥➏➍➃❸➋♠➌✞➏❛➂❩➆✥➏ ✛✻➯■✛✽✫⑤➃❸➄❳➔✵➈✍➉❛➈➡➒❸➘ ➆✂➄❛➅❩➈✍➊➡➃❷➆✍➒✿➊✥➆❫➄❛➈✹➋✥➑✿➎❁➈✥➆✥➎❼➒✇➃❸➌✚➈☎➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❁➻ ✖➋♠➌✚➄➍➃❸➎❁➈❄➉✆➪❳➂➇➆❫➏✵➂➇➆❫➅✚➅❩➈✍➌✚➄☎➪❳➂❁➈✍➌❞➪✖➈☎➐✚➄❛➈✪➏❛➂✚➈✘✛✻➯✢✛✽✫ ➷❤➐❁➈✍➐✚➈➡➓ ➃❸➌✚Ð✫➎❼➃❸➄❛➊➡➃❸➅❤➒✇➃❷➌❁➈❑➃❸➌❁➄❛➏❛➈★➆❫➎✽➋❫➑✏➼✛✂✯➻✂➸➴➋✥➪ ➪➴➂✚➈✍➌✽➆✸➶✚➐✚➉❛➄❛➏ ➑❦➉❛➋❭➔ ➋❭➌❁➈✏➄❛➋♠➐✚➉❛➊✍➈✏➆❫➉❛➉➍➃❸➚❭➈✍➄✍➹✖➏❛➂❤➃❸➄✵➶✚➐❁➉❛➄❛➏✵➅➇➆❫➄❛➄❛➈❄➄✵➏❛➂❁➉❛➋❭➐❁Ð❭➂❞➏❛➂❁➈☎➷❤➐✚➈✍➐❁➈☎➃❸➌ ➆✞➊➡➒❸➐❁➔✲➅❑➪➴➂❤➃✇➒❸➈✞➄❛➐❁➶✚➄❛➈✍➷❤➐✚➈✍➌❤➏❋➅❩➆✥➊➡✃❭➈✍➏❛➄✵➑❦➉❛➋♠➔ ➏❛➂✚➈❋➋♠➏❛➂✚➈✍➉☎➄❛➋♠➐✚➉❛➊✍➈✍➄ ➆✥➉❛➈✹➏❛➈✍➔✵➅❩➋❭➉r➆❫➉➍➃✇➒❸➘✞➎❁➈r➒❷➆★➘❭➈✍➎✍✰☛➏❛➂❼➃❸➄❳➒❷➆❫➏❛➏❛➈✍➉Ò➎❁➈➡➒❷➆✍➘♠➹☛➂✚➋✥➪✖➈✍➚❭➈✍➉✍➹❁➃❸➄Ò➔✯➐✚➊➡➂ ➄❛➔➾➆★➒✇➒❸➈✍➉✹➏❛➂➇➆❫➌❋➏❛➂✚➈✹➎❁➈r➒❷➆★➘✏➏❛➂❩➆✥➏✠➏❛➂✚➈✹➶❁➐✚➉❛➄❛➏➍➃❸➌✚Ð✪➄❛➋♠➐✚➉❛➊✍➈✂➪✖➋♠➐❤➒❸➎❋➂➇➆★➚❭➈ ➉❛➈❄➊✍➈➡➃❸➚♠➈❄➎✸➐✚➌❁➎✚➈✍➉☎➼✛✂✯➻❢➁❳➂❤➐❁➄✍➹✦➏❛➂❁➈✏➅❼➒❷➆✍➘♠➓❏➶❩➆✥➊➡✃→➅❩➋❫➃❸➌❤➏☎➌✚➈✍➈✍➎❞➌✚➋♠➏ ➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð ➔✵➈✥➆✥➌ ✓✬✓❼➻ ✓✂✁P➃✇➒❸➈ ➼✛✂ ✝❼➻✒✂✔✙ ✗✖✝❼➻ ✧✖✙ ✛✻➯✢✛✽✫ ✝❼➻✒✂ ✣ ✝✲✟❁➻ ✣ ✮ ➁✿➆❫➶❤➒❸➈ ✂♠❒✂➁❳➂❁➈✲➔✵➈✥➆❫➌①➆❫➌✚➎✫✓✖✓❤➻✼✓✻✳ ➏❛➂✞➅❩➈✍➉❛➊✍➈❄➌❤➏➍➃✇➒❸➈✪➷❤➐✚➈✍➐❁➈r➃❸➌❁Ð▼➎✚➈➡➒❷➆✍➘❤➄ ➭❦➔✵➈✥➆✥➄❛➐❁➉❛➈❄➎Ý➃❸➌✸➔✂➃✇➒✇➒✇➃❸➄❛➈✍➊✍➋♠➌➇➎❁➄✍➹✹➪❳➂❤➃❸➊➡➂❞➃❸➄✏➆ ➄➍➃❸➌✚Ð❫➒❸➈✞➅❩➆✥➊➡✃❭➈✍➏☎➏❛➉r➆❫➌✚➄➍➓ ➔✯➃❸➄❛➄➍➃❸➋❭➌▼➏➍➃❸➔✵➈✥➺✖➑❦➋♠➉➮➆➾➄r➆✥➔✵➅❤➒❸➈✯➱❩➋✥➪✛➐✚➌✚➎❁➈✍➉✹➏❛➂✚➈✂➼✛✂ ➆❫➌✚➎✪✛✻➯■✛✽✫ ➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✞➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✵➄✍➻❳➁❳➂✚➈✠➒✇➃❸➌❁✃✪➃❸➄ ✧✖✝❤➻✼✗✄✁✔➐❁➏➍➃✇➒✇➃❸à❄➈✍➎✧➻ ➶❩➈❞➔✵➋✥➚♠➈❄➎✛➋♠➐✚➏✞➆❫➄❋➑❏➆❫➉❋➏❛➋ ➆❫➊✍➊❄➋♠➔✵➔✵➋❤➎➇➆❫➏❛➈→➏❛➂❁➈❋✲➍➃❸➏❛➏❛➈❄➉✪➃❸➌❁➎✚➐❁➊❄➈✍➎ ➶❤➘▼➏❛➂✚➈❋➶❁➐✚➉❛➄❛➏✍➻✚✛❁➐✚➉❛➏❛➂❁➈❄➉❛➔✵➋♠➉❛➈♠➹❳➏❛➂✚➈✪➅➇➆❫➉❛➏➍➃❸➊❄➐❼➒❷➆✥➉✪➄❛➋♠➐✚➉❛➊✍➈❋➅✚➉❛➋❤➎❁➐✚➊➡➓ ➃❸➌✚Ð❋➏❛➂❁➈✵➶✚➐✚➉❛➄❛➏P➃❸➄✹➌✚➋♠➏✹➄➍➃❸➌✚Ð❫➒❸➈✍➎❞➋❭➐❁➏✹➑❦➋♠➉❳➃❸➌❁➊❄➉❛➈✥➆❫➄❛➈✍➎❋✲➍➃❸➏❛➏❛➈❄➉☛✰✿➆★➒✇➒✦➏❛➂❁➈ ➄❛➋♠➐✚➉❛➊✍➈✍➄✦➄❛➂❩➆✥➉❛➈Ô➃❸➌➾➆★➒✇➒✰➏❛➂❁➈✦✲➍➃❸➏❛➏❛➈✍➉❢➃❸➌❁➎✚➐❁➊❄➈✍➎✪➶❤➘✂➏❛➂❁➈❳➶✚➐❁➉❛➄❛➏❛➄✿➋❫➑❱➆✍➒✇➒☛➏❛➂❁➈ ➄❛➋♠➐✚➉❛➊✍➈✍➄✍➻ ✬✖➈✍➊✥➆✍➒✇➒❳➏❛➂❩➆✥➏✹➪❳➂❁➈✍➌⑧➏❛➂❁➈✵➅➇➆❫➊r✃♠➈❄➏❛➄✵➆✥➉❛➈✵➋❫➑❳➐✚➌❤➃✇➑❦➋♠➉❛➔ ➄➍➃❸à✍➈♠➹ ➏❛➂✚➈✵➏❛➋♠➏r➆✍➒✖➷❤➐✚➈✍➐✚➈➡➃❸➌❁Ð✞➎✚➈➡➒❷➆✍➘ ➃❸➌⑧➆❫➌❤➘✞➏➍➃❸➔✵➈✵➅❩➈❄➉➍➃❸➋❤➎✸➭❦➄❛➐❁➔✲➔✵➈✍➎▼➋✥➚♠➈❄➉ ➆★➒✇➒✦➱❩➋✥➪❳➄r➺➴➃❸➄P➃❸➌✚➎✚➈✍➅❩➈✍➌✚➎❁➈❄➌❤➏✆➋✥➑❳➏❛➂✚➈✆➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð❞➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✪➻❋➁❳➂❁➈ ✛✻➯✢✛✦✫✔➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔❐➄❛➅❼➒✇➃❸➏❛➄✂➏❛➂❤➃❸➄✹➎❁➈➡➒❷➆✍➘❋➈✍➚❭➈✍➌❤➒❸➘♠➹✻➪❳➂✚➈✍➉❛➈✥➆✥➄✹➏❛➂❁➈✂➼✛✂ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➆❫➄❛➄➍➃❸Ð❭➌❁➄❳➏❛➂✚➈➴➎✚➈➡➒❷➆✍➘✵➏❛➋✹➏❛➂✚➈➴➱➇➋✥➪➴➄✿➏❛➂➇➆❫➏✿➊✥➆❫➐✚➄❛➈✍➎☎➏❛➂❁➈❳➔✵➋✥➓ ➔✵➈✍➌❤➏r➆✥➉❛➘✵➷❤➐✚➈✍➐❁➈r➃❸➌❁Ð✪➭❦➶❤➘✂➄❛➈✍➌✚➎❼➃❸➌✚Ð✵➶✚➐❁➉❛➄❛➏❛➄r➺➡➻✿➼→➂❁➈✍➌☎➏❛➂✚➈❳➎❁➈➡➒❷➆✍➘❤➄Ò➆❫➉❛➈ ➄❛➂➇➆❫➉❛➈✍➎⑧➆❫➄✠➃❸➌✤✛✻➯✢✛✦✫✯➹❁➃❸➌▼➪❳➂➇➆❫➏✠➔✂➃❸Ð♠➂❤➏✹➶❩➈✵➊★➆★➒✇➒❸➈✍➎❞➆✵➔✂➐❼➒❸➏➍➃❸➅❤➒❸➈✍➬♠➃❷➌❁Ð ➋❫➑✰➶✚➐❁➉❛➄❛➏❛➄✍➹✚➏❛➂✚➈✖Ï❩➫♠➥❄↕ ✟❄➝❤➤❫↕❦➫✍✲➍➃❸➏❛➏❛➈✍➉✿➶❩➋♠➐✚➌❁➎✚➄Ò➆❫➉❛➈❳➄❛➔☎➆✍➒✇➒❸➈✍➉❳➏❛➂❩➆✥➌✆➪❳➂✚➈✍➌ ➏❛➂✚➈✆➄❛➋❭➐❁➉❛➊❄➈✍➄☎➆❫➉❛➈✵➃❸➄❛➋✥➒❷➆❫➏❛➈✍➎✸➑❦➉❛➋❭➔❅➈✥➆✥➊➡➂❞➋♠➏❛➂✚➈✍➉✲➆❫➄✯➃❸➌❞➼✛✂✯➻✰➁➴➂❤➃❸➄ ➪❳➆✥➄✖➈✍➬✚➆❫➊❄➏➍➒❸➘✪➋♠➐✚➉❳Ð♠➋ä➆★➒✆✰☛➐✚➌❁➎✚➈✍➉❳➏❛➂❁➈✹➄r➆✥➔✵➈➴➒✇➃❸➌✚✃❋➐✚➏➍➃✇➒✇➃❸à✥➆❫➏➍➃❸➋❭➌✧➹❋✛✻➯■✛✽✫ ➆★➒✇➒❸➋✥➪❳➄✦➆➴➌❤➐✚➔✯➶❩➈❄➉❢➋❫➑☛➄❛➋❭➐❁➉❛➊✍➈❄➄✿➆❫Ð♠Ð❭➉❛➈✍Ð❤➆✥➏➍➃❸➌❁Ð✂➏❛➂✚➈➡➃❸➉❢➏❛➉r➆★Ó☎➊✦➏❛➋✠➋♠➶✚➏r➆★➃❸➌ ➆P➒❸➋✥➪✖➈✍➉❳➋✥➚♠➈❄➉r➆★➒✇➒✿➎✚➈➡➒❷➆✍➘✿✲➍➃❸➏❛➏❛➈✍➉✍➻ ➯➍➌❞➋♠➉❛➎✚➈✍➉☎➏❛➋Ñ➏❛➈❄➄❛➏✯➋❭➐❁➉✵➃❸➌❤➏❛➐❤➃❸➏➍➃❸➋♠➌☛➹✠➪✖➈✏➂❩➆✍➚❭➈✪➄➍➃❸➔✂➐❼➒❷➆✥➏❛➈✍➎❑➶❩➋♠➏❛➂ ➏❛➂✚➈✵➼✛✂✔➆❫➌✚➎✤✛✻➯✢✛✽✫ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄✍➻✵➁❳➂✚➈✠☎✠➅✚➅❩➈✍➌✚➎❼➃❸➬✞➊✍➋❭➌❤➏r➆★➃❸➌✚➄✵➆ ➊✍➋❭➔✵➅❼➒❸➈❄➏❛➈✪➎✚➈✍➄❛➊✍➉➍➃❸➅✚➏➍➃❸➋♠➌❑➋❫➑Ô➋♠➐✚➉✯➄➍➃❸➔✂➐❼➒❷➆✥➏➍➃❸➋♠➌❑➅✚➉❛➋❤➊✍➈✍➎✚➐❁➉❛➈❏✰✖➪✖➈☎➋♠➌❤➒❸➘ ➅✚➉❛➈✍➄❛➈✍➌❤➏✵➏❛➂✚➈✪➉❛➈✍➄❛➐❤➒❸➏❛➄☎➂❁➈❄➉❛➈♠➻✞➼Ñ➈✏➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉✪➆❋➄➍➃❸➌✚Ð❫➒❸➈✪✂ â✪➶❼➃❸➏✆☎✥➄❛➈✍➊ ➒✇➃❸➌✚✃✪➶❩➈➡➃❸➌✚Ð☎➐❁➏➍➃✇➒✇➃❸à❄➈✍➎Ñ➶❤➘ ✂✭✕P➱❩➋✥➪❳➄✍➹✚➈✥➆❫➊r➂✪➂➇➆✍➚♠➃❸➌❁Ð☎➏❛➂✚➈✹➄r➆❫➔✵➈Ò➄❛➏r➆❫➏➍➃❸➄➍➓ ➏➍➃❸➊✥➆✍➒✿Ð♠➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌Ñ➅✚➉❛➋❤➊✍➈✍➄❛➄✍➻Ò➯➍➌✏➁✿➆❫➶❤➒❸➈ ✂✂➪✖➈✹➄❛➂✚➋✥➪✛➏❛➂✚➈✹➔✵➈✥➆❫➌✞➆✥➌❁➎ ✓✖✓❤➻✼✓✻✳ ➏❛➂❑➅➇➈✍➉❛➊✍➈✍➌❤➏➍➃✇➒❸➈❞➷❤➐✚➈✍➐❁➈r➃❸➌❁Ð✫➎❁➈r➒❷➆★➘❤➄✪➑s➋♠➉❋➆▼➄r➆✥➔✵➅❼➒❸➈▼➱➇➋✥➪ ➭❦➏❛➂❁➈ ➎➇➆❫➏r➆✆➑❦➉❛➋❭➔ ➏❛➂✚➈☎➚❭➆❫➉➍➃❸➋♠➐✚➄✯➱➇➋✥➪➴➄✲➆❫➉❛➈☎➄➍➃❸➔✯➃✇➒❷➆✥➉r➺✵➐❁➌✚➎❁➈❄➉✯➈✥➆✥➊➡➂⑧➋❫➑Ò➏❛➂❁➈ ➏➍➪✦➋❋➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❑➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✵➄✍➻→➸❳➋♠➏❛➈✪➏❛➂➇➆❫➏✲➪➴➂❤➃✇➒❸➈✪➏❛➂✚➈✪➔✲➈✥➆❫➌❞➎✚➈➡➓ ➒❷➆✍➘❤➄✹➆✥➉❛➈✵➆✥➶❩➋♠➐✚➏✠➏❛➂✚➈✹➄r➆❫➔✵➈✹➑❦➋❭➉➴➏❛➂✚➈✂➏➍➪✖➋✲➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄✍➹❢➏❛➂✚➈✵✓✖✓❼➻ ✓ ✳ ➏❛➂ ➅❩➈✍➉❛➊❄➈✍➌❤➏➍➃✇➒❸➈Ñ➎✚➈➡➒❷➆✍➘❤➄✏➆❫➉❛➈❋➄➍➃❸Ð❭➌❼➃✇➽➇➊✥➆❫➌❤➏➍➒❸➘❑➄❛➔➾➆★➒✇➒❸➈✍➉✏➐❁➌✚➎❁➈❄➉✵➏❛➂❁➈✞✛✻➯■✛✽✫ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✪➻❳➁❳➂❼➃❸➄Ò➊✍➋♠➌❤➽➇➉❛➔✵➄✠➋❭➐❁➉Ò➆❫➌➇➆★➒❸➘❤➄➍➃❸➄✂➆✥➶❩➋✥➚♠➈♠➻ ➁➴➂✚➈✞✛✻➯✢✛✦✫❅➷❤➐❁➈✍➐✚➈✞➎❼➃❸➄❛➊➡➃❸➅❤➒✇➃❷➌❁➈❞➂➇➆❫➄✪Ð❭➈✍➌✚➈✍➉r➆★➒✇➒❸➘✫➶❩➈✍➈✍➌❑➊❄➋♠➌✚➄➍➃❸➎❼➓ ➈✍➉❛➈✍➎✸➃❸➌✚➈➡Þ☛➈❄➊✍➏➍➃❸➚♠➈▼➑❦➋❭➉✆➅✚➉❛➋✥➚♠➃❸➎❼➃❸➌✚Ð❑➉❛➈✥➆✍➒✇➓➍➏➍➃❸➔✲➈▼➄❛➈✍➉❛➚♠➃❸➊✍➈✒✰✹➃❸➌✸➑❏➆✥➊✍➏✍➹❳➃❸➏ ➂➇➆❫➄✪➶➇➈✍➈✍➌❑➄❛➂✚➋❫➪❳➌❞➃❸➌❑➊✍➈✍➉❛➏r➆★➃❸➌✫➊➡➃❸➉❛➊✍➐✚➔✵➄❛➏r➆✥➌❁➊✍➈❄➄❋➏❛➋▼➶❩➈❋➏❛➂✚➈✸➩❢➫❤➜❛➥✍↕ ➅❩➋❭➄❛➄➍➃❸➶❼➒❸➈✲➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➭✢✍✒✂ ✎s➺➡➻❢➁❳➂❁➈Ò➉❛➈✥➆❫➄❛➋❭➌✆➃❸➄✖➏❛➂➇➆❫➏✿➃✇➑✿➋♠➌✚➈✠➄❛➋♠➐✚➉❛➊✍➈✠➃❸➌❤➓ ✲❛➈✍➊❄➏❛➄✠➈✍➬❤➊✍➈❄➄❛➄➍➃❸➚♠➈☎➏❛➉r➆★Ó☎➊✂➃❸➌❤➏❛➋✪➏❛➂✚➈✯➌✚➈✍➏✍➹✰➏❛➂❼➃❸➄✹➎❤➃❸➄❛➉❛➐❁➅✚➏❛➄✂➏❛➂❁➈✵➄❛➈❄➉❛➚♠➃❸➊✍➈ ➑❦➋❭➉❢➈✍➚♠➈✍➉❛➘❭➋♠➌✚➈♠➻✿➁❳➂❼➃❸➄✦➆❫➄❛➄❛➈✍➄❛➄❛➔✲➈✍➌❤➏✍➹❤➂✚➋❫➪✦➈✍➚♠➈✍➉✍➹❼➆❫➉➍➃❸➄❛➈✍➄❢➑❦➉❛➋♠➔ ➆➴➑❏➆★➃✇➒❸➐✚➉❛➈ ➏❛➋✂➎❼➃❸➄❛➏➍➃❸➌✚Ð♠➐❤➃❸➄❛➂Ñ➏❛➂✚➈✠➏❏➪✖➋✹➄❛➈✍➅➇➆❫➉r➆✥➏❛➈✠➋♠➶✻✲❛➈❄➊✍➏➍➃❸➚♠➈✍➄Ò➋❫➑❱➆❫➌❤➘✲➏❛➉r➆★Ó☎➊Ò➊✍➋♠➌❤➓ ➏❛➉❛➋❫➒✦➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✪➹Ô➢ ➥✍➫✖✥❸➝❤↕❦➢③➫❤↔✲➆❫➌✚➎Ý➥■✦❩➝❤➜r➢③↔❤➛♠➻✖➯❏➄❛➋❫➒❷➆❫➏➍➃❸➋❭➌▼➃❸➄✹➏❛➂❁➈✂➔✵➋♠➉❛➈ ➑❦➐✚➌❁➎➇➆❫➔✲➈✍➌❤➏r➆★➒➮Ð♠➋❤➆✍➒✆✰✠➃❸➏☎➅❁➉❛➋✥➚♠➃❸➎✚➈✍➄✪Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎❑➄❛➈✍➉❛➚♠➃❸➊✍➈✪➑❦➋❭➉✵➪✖➈➡➒✇➒✇➓ ➶❩➈✍➂➇➆✍➚♠➈✍➎⑧➊➡➒✇➃❸➈✍➌❤➏❛➄☎➆✥➌❁➎✞➷❤➐❩➆✥➉r➆❫➌❤➏➍➃❸➌✚➈✍➄✵➔✯➃❸➄❛➶➇➈✍➂❩➆✍➚♠➃❸➌✚Ð▼➄❛➋♠➐✚➉❛➊✍➈✍➄✍➻ ✎✳➐✚➏ ➄❛➂➇➆❫➉➍➃❸➌✚Ð❁➹❁➃✇➑☛➃❸➏✳➃❸➄✦➅❩➈✍➉➍➑❦➋❭➉❛➔✵➈✍➎✯➃❸➌☎➏❛➂✚➈✠➊✍➋❭➌❤➏❛➈✍➬❤➏✳➋✥➑❱➆❫➌☎➈✍➌✚➊✍➋♠➔✵➅➇➆❫➄❛➄➍➃❸➌✚Ð ➃❸➄❛➋❫➒❷➆✥➏➍➃❸➋♠➌▼➄❛➊r➂✚➈✍➔✵➈♠➹➇➅❩➈✍➉➍➑❦➋❭➉❛➔✵➄❳➏❛➂❁➈✹➚❭➈✍➉❛➘☎➎❼➃✇Þ☛➈❄➉❛➈✍➌❤➏✹Ð♠➋ä➆★➒❢➋✥➑❢➔✯➃❸➬♠➃❸➌✚Ð ➏❛➉r➆★Ó➾➊➾➑s➉❛➋♠➔❐➎❼➃✇Þ❙➈✍➉❛➈✍➌❤➏✵➄❛➋❭➐❁➉❛➊✍➈❄➄✯➃❸➌→➆✪➪❳➆✍➘Ñ➏❛➂➇➆❫➏✹➃❸➄✵➶❩➈✍➌✚➈➡➽➇➊➡➃❷➆★➒Ò➏❛➋ ➆★➒✇➒✆✰➮➶❁➐✚➉❛➄❛➏❛➄❋➆✥➉❛➈❋➔✯➐❤➒❸➏➍➃❸➅❼➒❸➈❄➬❤➈✍➎✛➄❛➋▼➏❛➂➇➆❫➏☎➏❛➂✚➈✪Ï❩➫♠➥❄↕ ✟✍➝❤➤✥↕❦➫ ✲❏➃❸➏❛➏❛➈✍➉✵➃❸➄ ➄❛➔☎➆✍➒✇➒❸➈✍➉➴➑❦➋❭➉✖➈✍➚❭➈✍➉❛➘♠➋❭➌❁➈♠➻✿➁❳➂✚➈✌✛✻➯✢✛✦✫⑤➄❛➊r➂❁➈✍➔✲➈➴➃❸➄Ò➆❫➌☎➈➡Þ☛➈❄➊✍➏➍➃❸➚♠➈✂➄❛➂❩➆✥➉➍➓ ➃❸➌✚Ð✵➄❛➊➡➂✚➈✍➔✵➈❭➹❁➶✚➐✚➏Ô➃❸➏✦➎❁➋❤➈✍➄❳➌✚➋♠➏✖➅✚➉❛➋✥➚♠➃❸➎❁➈✲➆❫➌❤➘✵➃❸➄❛➋✥➒❷➆❫➏➍➃❸➋♠➌☛➻❳➼✛✂✯➹✚➋♠➌ ➏❛➂✚➈✆➋❭➏❛➂❁➈❄➉✯➂➇➆❫➌✚➎✧➹❢➃❸➄✲➆❫➌❞➈➡Þ☛➈❄➊✍➏➍➃❸➚♠➈✏➔✵➈✍➏❛➂✚➋❤➎Ñ➑❦➋♠➉✹➃❸➄❛➋❫➒❷➆✥➏➍➃❸➋♠➌☛➻✞➯◗➑Ò➪✖➈ ➋♠➉❛Ðä➆❫➌❤➃❸à✍➈✠➏❛➂✚➈❳➏❛➉r➆★Ó☎➊✖➃❸➌❤➏❛➋✹➊➡➒❷➆✥➄❛➄❛➈✍➄✖➋✥➑✻➊r➒✇➃❸➈✍➌❤➏❛➄✖➪✖➃❸➏❛➂✵➄➍➃❸➔✯➃✇➒❷➆✥➉➴➄❛➈❄➉❛➚♠➃❸➊✍➈ ➉❛➈✍➷❤➐❤➃❸➉❛➈✍➔✵➈✍➌❤➏❛➄✍➹❱➪✖➈P➽➇➌✚➎Ñ➏❛➂➇➆❫➏✹➏❛➂❤➃❸➄✹➉❛➈✥➆❫➄❛➋❭➌❼➃❸➌✚Ð✪➒❸➈✥➆✥➎❁➄✂➏❛➋✏➆➾➌✚➈✍➄❛➏❛➈✍➎ ➄❛➊r➂❁➈✍➔✲➈➴➃❸➌✪➪❳➂❤➃❸➊➡➂✏➏❛➂❁➈Ò➷❤➐❁➈❄➐❼➃❸➌✚Ð☎➎❁➈✍➊r➃❸➄➍➃❸➋♠➌✪➃❸➄✖➃❸➌✪➏❏➪✖➋✹➄❛➏❛➈✍➅✚➄✍❒✖➆✠➽➇➉❛➄❛➏ ➄❛➏❛➈✍➅✂➏❛➋➴➃❸➌✚➄❛➐❁➉❛➈✖➃❸➄❛➋✥➒❷➆❫➏➍➃❸➋♠➌✏➋❫➑☛➊r➒❷➆❫➄❛➄❛➈✍➄✍➹✚➆✥➌❁➎✂➏❛➂✚➈✍➌☎➆➴➅➇➆❫➉❛➏➍➃❸➊❄➐❼➒❷➆✥➉✖➄❛➂❩➆✥➉➍➓ ➃❸➌✚Ð✆➔✲➈✍➏❛➂❁➋❤➎❋➪✖➃❸➏❛➂❤➃❸➌❋➈✥➆❫➊r➂❋➊➡➒❷➆✥➄❛➄✍➻ ✛✻➯✢✛✽✫✔➃❸➄✵➌✚➋♠➏✵➏❛➂✚➈✪➋❭➌❼➒❸➘▼➃❸➌❤➏❛➈✍➉❛➈❄➄❛➏➍➃❸➌❁Ð❞➄❛➂➇➆❫➉➍➃❸➌✚Ð▼➔✵➈✍➏❛➂✚➋❤➎✧➻✤☎✠➌❤➓ ➋♠➏❛➂✚➈✍➉Ò➄❛➂❩➆✥➉➍➃❸➌❁Ð✏➔✵➈✍➏❛➂✚➋❤➎✪➃❸➄✠➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘❭➹✻➪❳➂❤➃❸➊➡➂❋➂➇➆❫➄➮➆✯➚♠➈✍➉❛➘✏➎❼➃✇Þ❙➈✍➉❛➈✍➌❤➏ ➶❩➈✍➂➇➆✍➚♠➃❸➋♠➉❳➏❛➂➇➆❫➌✆✛✻➯■✛✽✫✯➻❭➁➴➂✚➈❳Ð♠➋ä➆★➒☛➋✥➑✯✛✻➯■✛✽✫Ý➃❸➄✖➏❛➋✠➒❸➈❄➏✳➈✍➚❭➈✍➉❛➘✂➄❛➋♠➐✚➉❛➊✍➈
Host-1 Host-2 Host-3 Host-4 Host-5 s-3 S- Figure Q Net work topology used for data in Table 2. Path length FIFO+ fowl d cQ fY2Y Table 2: The mean and ,,'th percentile queueing de lays(measured in milliseconds, which is a single packet transi time)for four sample flows of different path leng ths under the WfQ, FIFO, and FIFO+ scheduling algorit hms. The network configuration is shown in Figure Q Each inter-switch link is 8w Yyo utilized in a common class share equally in the jitter. In priority, jitter for all sources in the aggregate class) across all the hops class acquires jitter of higher priority classes, which con- along the path to minimize jitter. We do this as follows. For lently get much lower jitter. In one direction priority is each hop we measure the average delay seen by packets in idered a sharing mechanism, but in the ot her it is each aggregate class at that switch. We then compute for lation mechanism, i.e. lower priority traffic can never affect each packet the difference between its particular delay and Why might a priority algorithm be of mutual benefit? a field in the header of the packet, which thus accumulate he benefit of lower jitter is obvious; the benefit of higher the total offset for this packet from the average for its clas jitter would presumably be a lower cost for the service This field allows each switch to compute when the packet e with more tolerance for jitter (or for hig her shou have arrived if it were indeed given average service delay) might be very happy to obtai cheaper service in The switch then schedules the packet as if it arrived at exchange for taking the jitter of ot average expected time; this is done by ordering the queue resenting methods for jitter shirting, in which expuiciaction One can think in general of scheduling algorithms as re by these expected arrival times rather than by the act ual arrival tim are taken to transfer the jitter among flows in a controlled To test this algorithm, we have simulated its performance nd characterized way. One could invent a wide range of on a net work as shown on Figure Q This net work has four that reorder th uivalent Q Mbit/sec inter-switch links, and each link is as we discuss in the section lated work. They should shared by c flows. There are, in total, 22 flows; all of all be examined from two perspectives. First, how and to them have the same statistical generation process(described hat extent do they perform isolation? Second, how and to opendix) but they travel different net work paths what extent do they provide sharing Qe traverse only inter-switch link, f traverse two inter- witch links, f links, and 2 5 V ulti-Hop sharinA verse all four inter-switch links. Table 2 displays the mean and ,,,'th percentile queueing delays for a single sample One of the problems with the FiFO algorithm is that if flow for each path leng th (the data from the alize our gedank t to include ver similar). We compare the WFQ, FIFO, and FIFO+ alge links, then the jitter tends to increase dramatically with the rithms (where we have used equal clock rates in the WFQ number of hops, since the packet has a se parate opportunit Igorithm). Note that the mean delays are comparable in for uncorre lated queueing delays at each hop all three cases. While the,,,'th percentile delays increase In fact, it is not clear that this increase in jit ter need with path length for all three algorithms, the rate of growth occur. Poing through more hops provides more opportuni is much smaller with the Fifo++ algorith ities for red g s the simulation shows, the effect of FIFO+, ed to FIFO, is to slig ht ly inc he delay and jit. flows on short paths, slightly de cheme FIFO+.FIFO+ ithms for manufacturing systems ificantly de the jitter of fows on long paths, east sack scheduling al ns that th erall de lay bound goes down and discussed in Reference [ CSE of est imation goes up on long paths. When In FIFO+, we try to induce FIFO-style sharing (equal mplementation of the two schemes, they differs
Host-1 Host-2 Host-3 Host-4 Host-5 S-1 S-2 S-3 S-4 S-5 ✛✧➃❸Ð♠➐✚➉❛➈ ✂❭❒✦➸➴➈✍➏❏➪✖➋♠➉❛✃☎➏❛➋♠➅➇➋❫➒❸➋♠Ð❭➘❋➐❁➄❛➈❄➎✪➑❦➋♠➉✖➎➇➆❫➏r➆✯➃❸➌❋➁✿➆✥➶❼➒❸➈❉✮❤➻ ➵❢➆✥➏❛➂ ✳☛➈✍➌❁Ð❭➏❛➂ ✂ ✮ ✝ ✟ ➄❛➊r➂❁➈❄➎❁➐❤➒✇➃❸➌❁Ð ➔✵➈✥➆❫➌ ✓✖✓❤➻✼✓ ✁P➃✇➒❸➈ ➔✵➈✥➆❫➌ ✓✖✓❼➻ ✓✂✁P➃✇➒❸➈ ➔✵➈★➆❫➌ ✓✬✓❼➻ ✓✂✁P➃✇➒❸➈ ➔✵➈✥➆✥➌ ✓✬✓❼➻ ✓✂✁P➃✇➒❸➈ ➼✛✂ ✮❼➻ ✙✖✗ ✟❨✗❼➻ ✝✾✂ ✟✚➻ ✣✟ ✙✖✕❤➻✼✝●✂ ✣ ➻ ✗✾✂ ✙✬✗❼➻ ✧✖✙ ✓❤➻✼✙✔✟ ✧✬✕❼➻ ✗✖✓ ✛✻➯■✛✽✫ ✮❼➻ ✗✲✟ ✝✬✕❼➻ ✟●✓ ✟✚➻ ✣ ✝ ✟✻✂❭➻ ✮❏✮ ✣ ➻ ✓ ✣ ✗❏✮❼➻ ✝✖✙ ✂✭✕❼➻ ✝✖✝ ✗✬✧❼➻✒✂✔✝ ✛✻➯■✛✽✫ ✰ ✮❼➻ ✣ ✂ ✝✬✝❼➻ ✗✖✓ ✟✚➻✼✙✬✓ ✝✖✧❤➻✰✂✭✗ ✣ ➻ ✣ ✙ ✟❨✝❼➻ ✝✖✕ ✂✭✕❼➻✒✂✖✂ ✟❨✗❼➻ ✮✖✗ ➁✿➆✥➶❼➒❸➈ ✮❼❒✹➁❳➂✚➈✯➔✵➈★➆❫➌❞➆✥➌❁➎✤✓✖✓❤➻✼✓✻✳ ➏❛➂✞➅❩➈✍➉❛➊✍➈❄➌❤➏➍➃✇➒❸➈✆➷❤➐✚➈✍➐✚➈➡➃❸➌✚ÐÑ➎✚➈➡➒❷➆✍➘❤➄✆➭❦➔✲➈✥➆❫➄❛➐✚➉❛➈✍➎Ñ➃❸➌▼➔✂➃✇➒✇➒✇➃❸➄❛➈✍➊✍➋♠➌✚➎✚➄✍➹✖➪❳➂❼➃❸➊r➂Ñ➃❸➄✂➆☎➄➍➃❸➌❁Ð✥➒❸➈✆➅➇➆❫➊r✃♠➈❄➏✯➏❛➉r➆✥➌❁➄❛➔✯➃❸➄❛➄➍➃❸➋❭➌ ➏➍➃❸➔✲➈✥➺✳➑❦➋♠➉✳➑s➋♠➐✚➉✖➄r➆❫➔✵➅❤➒❸➈✹➱❩➋✥➪❳➄✖➋❫➑❱➎❼➃✇Þ❙➈✍➉❛➈✍➌❤➏✠➅➇➆❫➏❛➂☎➒❸➈✍➌✚Ð♠➏❛➂✚➄✠➐✚➌❁➎✚➈✍➉❳➏❛➂❁➈Ò➼✛✂✯➹❉✛✻➯■✛✽✫✯➹➇➆❫➌✚➎ ✛✻➯■✛✽✫ ✰ ➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð❋➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✵➄✍➻❳➁❳➂❁➈Ò➌❁➈❄➏➍➪✖➋❭➉❛✃ ➊❄➋♠➌❤➽❩Ð❭➐❁➉r➆✥➏➍➃❸➋♠➌Ñ➃❸➄❳➄❛➂✚➋❫➪❳➌✪➃❸➌✞✛✧➃❸Ð❭➐❁➉❛➈✆✂♠➻ ✎✿➆❫➊r➂✆➃❸➌❤➏❛➈✍➉➍➓❏➄❛➪✖➃❸➏❛➊➡➂❋➒✇➃❸➌✚✃✆➃❸➄✌✧✬✝❼➻ ✗ ✁✔➐✚➏➍➃✇➒✇➃❸à✍➈✍➎☛➻ ➃❸➌→➆✪➊✍➋❭➔✵➔✵➋❭➌Ý➊r➒❷➆❫➄❛➄✵➄❛➂➇➆❫➉❛➈✏➈✍➷❤➐❩➆✍➒✇➒❸➘Ý➃❸➌❞➏❛➂✚➈ ✲❏➃❸➏❛➏❛➈✍➉✍➻✞➯➍➌▼➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘♠➹ ➋❭➌❁➈❳➊➡➒❷➆✥➄❛➄✖➆✥➊✍➷❤➐❤➃❸➉❛➈✍➄✽✲➍➃❸➏❛➏❛➈✍➉✿➋❫➑❙➂❼➃❸Ð❭➂❁➈✍➉✦➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✵➊➡➒❷➆✥➄❛➄❛➈✍➄✍➹✚➪❳➂❼➃❸➊r➂✵➊✍➋❭➌❼➓ ➄❛➈❄➷❤➐❁➈✍➌❤➏➍➒❸➘❋Ð❭➈✍➏✠➔✂➐✚➊➡➂✪➒❸➋✥➪✖➈✍➉✼✲➍➃❸➏❛➏❛➈❄➉✍➻❳➯➍➌✪➋❭➌❁➈✂➎❤➃❸➉❛➈✍➊✍➏➍➃❸➋♠➌✞➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘❋➃❸➄ ➊❄➋♠➌✚➄➍➃❸➎❁➈✍➉❛➈❄➎✪➆✖➄❛➂➇➆❫➉➍➃❸➌✚Ð✯➔✲➈✍➊➡➂➇➆❫➌❤➃❸➄❛➔✪➹❤➶✚➐✚➏✻➃❸➌✵➏❛➂❁➈✿➋♠➏❛➂✚➈✍➉✻➃❸➏✻➃❸➄✦➆❫➌✹➃❸➄❛➋❫➓ ➒❷➆✥➏➍➃❸➋♠➌✞➔✵➈✍➊r➂❩➆✥➌❼➃❸➄❛➔✏➹❩➃❦➻ ➈♠➻❢➒❸➋✥➪✖➈✍➉❳➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘❋➏❛➉r➆★Ó☎➊✹➊✥➆✥➌✪➌✚➈✍➚♠➈❄➉✹➆★Þ❙➈✍➊✍➏ ➏❛➂✚➈✹➅❩➈✍➉➍➑s➋♠➉❛➔☎➆✥➌❁➊✍➈✂➋❫➑✿➂❤➃❸Ð♠➂✚➈✍➉Ò➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘❋➋♠➌✚➈♠➻ ➼✸➂❤➘❞➔✯➃❸Ð❭➂❤➏✏➆ ➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘❑➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➶➇➈✪➋❫➑➮➔✯➐✚➏❛➐➇➆★➒Ò➶❩➈✍➌✚➈➡➽➇➏✄ ➁❳➂✚➈✵➶❩➈✍➌✚➈➡➽➇➏✯➋✥➑✳➒❸➋✥➪✖➈❄➉❂✲➍➃❸➏❛➏❛➈❄➉✠➃❸➄✹➋♠➶❤➚♠➃❸➋♠➐✚➄☛✰✿➏❛➂❁➈✲➶❩➈✍➌✚➈➡➽❩➏✂➋❫➑Ô➂❼➃❸Ð❭➂❁➈✍➉ ✲❏➃❸➏❛➏❛➈✍➉✹➪✖➋♠➐❤➒❸➎▼➅✚➉❛➈✍➄❛➐✚➔☎➆❫➶❤➒❸➘▼➶❩➈✏➆➾➒❸➋✥➪✖➈❄➉✹➊✍➋♠➄❛➏✠➑❦➋❭➉✹➏❛➂❁➈✲➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻✆☎ ➄❛➋❭➐❁➉❛➊❄➈✪➪✖➃❸➏❛➂▼➔✵➋❭➉❛➈✆➏❛➋✥➒❸➈✍➉r➆❫➌✚➊✍➈✪➑❦➋❭➉ ✲➍➃❸➏❛➏❛➈✍➉☎➭❦➋♠➉✹➑❦➋♠➉✂➂❤➃❸Ð♠➂✚➈✍➉✵➋✥➚♠➈✍➉r➆✍➒✇➒ ➎✚➈➡➒❷➆✍➘✚➺✯➔✯➃❸Ð❭➂❤➏✯➶➇➈✵➚♠➈✍➉❛➘✞➂➇➆❫➅✚➅❤➘▼➏❛➋✪➋❭➶❁➏r➆✍➃❸➌✸➆☎➊➡➂✚➈✥➆✥➅❩➈✍➉✂➄❛➈✍➉❛➚♠➃❸➊✍➈✵➃❸➌ ➈❄➬❤➊➡➂➇➆❫➌✚Ð♠➈✹➑❦➋♠➉❳➏r➆✥✃♠➃❸➌❁Ð✏➏❛➂❁➈ ✲➍➃❸➏❛➏❛➈❄➉➴➋✥➑✿➄❛➋♠➔✵➈✹➋❭➏❛➂❁➈❄➉➴➄❛➋❭➐❁➉❛➊❄➈✍➄✍➻ ✫✠➌✚➈➴➊★➆❫➌✂➏❛➂❼➃❸➌✚✃✯➃❸➌✵Ð❭➈✍➌❁➈❄➉r➆★➒☛➋✥➑☛➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð☎➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✵➄Ò➆❫➄❢➉❛➈❄➅❼➓ ➉❛➈❄➄❛➈✍➌❤➏➍➃❸➌❁Ð✹➔✲➈✍➏❛➂❁➋❤➎✚➄✻➑❦➋❭➉✻◆❄➢③↕❦↕◗➙✥➜✻➥■✦❩➢✟✍↕❦➢③↔❤➛♠➹❛➃❸➌✂➪➴➂❤➃❸➊r➂✹➈✍➬❤➅❼➒✇➃❸➊r➃❸➏❳➆❫➊✍➏➍➃❸➋❭➌❁➄ ➆✥➉❛➈✵➏r➆❫✃❭➈✍➌✞➏❛➋✆➏❛➉r➆✥➌❁➄➍➑s➈✍➉✹➏❛➂❁➈✾✲➍➃❸➏❛➏❛➈✍➉✂➆✥➔✵➋♠➌✚Ð✆➱➇➋✥➪❳➄➴➃❸➌❞➆✵➊✍➋❭➌❤➏❛➉❛➋❫➒✇➒❸➈✍➎ ➆✥➌❁➎→➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✍➈❄➎✸➪❳➆✍➘♠➻ ✫✠➌❁➈✏➊✍➋♠➐❤➒❸➎✸➃❸➌❤➚♠➈✍➌❤➏✏➆❋➪✖➃❸➎❁➈✏➉r➆❫➌✚Ð♠➈✪➋✥➑ ➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð❋➄❛➊➡➂✚➈✍➔✵➈✍➄Ò➏❛➂❩➆✥➏❳➉❛➈✍➋♠➉❛➎✚➈✍➉✠➏❛➂✚➈✂➷❤➐❁➈✍➐✚➈✹➃❸➌❋➄❛➅❩➈✍➊➡➃✇➽➇➊✵➪❳➆✍➘❤➄✍➹ ➆✥➄✹➪✖➈✵➎❤➃❸➄❛➊✍➐✚➄❛➄✯➃❸➌▼➏❛➂✚➈✵➄❛➈✍➊✍➏➍➃❸➋❭➌Ý➋❭➌Ñ➉❛➈r➒❷➆❫➏❛➈✍➎❞➪✖➋❭➉❛✃❩➻✵➁❳➂❁➈❄➘❋➄❛➂❁➋❭➐❼➒❸➎ ➆✍➒✇➒❳➶❩➈☎➈✍➬✚➆✥➔✯➃❸➌❁➈❄➎Ý➑❦➉❛➋❭➔ ➏❏➪✖➋✪➅➇➈✍➉❛➄❛➅❩➈✍➊✍➏➍➃❸➚❭➈✍➄✍➻✤✛✧➃❸➉❛➄❛➏✍➹✿➂❁➋✥➪✾➆✥➌❁➎⑧➏❛➋ ➪❳➂➇➆❫➏✖➈❄➬❤➏❛➈✍➌❤➏❳➎❁➋✂➏❛➂❁➈❄➘✵➅❩➈✍➉➍➑❦➋❭➉❛➔✔➃❸➄❛➋❫➒❷➆✥➏➍➃❸➋♠➌ Ý➲❤➈✍➊✍➋❭➌❁➎☛➹❩➂✚➋✥➪✫➆❫➌✚➎☎➏❛➋ ➪❳➂➇➆❫➏❳➈✍➬❤➏❛➈✍➌❤➏Ò➎❁➋✲➏❛➂❁➈✍➘☎➅✚➉❛➋✥➚♠➃❸➎❁➈✵➄❛➂➇➆❫➉➍➃❸➌✚Ð ✁❋Û✻❂ ❽❛Ü ❄✄✂✖Ù❉✸ ✏❁❚❤❿❤❾❦Ü❸Ø ☎ ✫✠➌✚➈✞➋❫➑✲➏❛➂❁➈▼➅✚➉❛➋♠➶❤➒❸➈✍➔✵➄✞➪✖➃❸➏❛➂❑➏❛➂❁➈ ✛✻➯✢✛✽✫ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➃❸➄❋➏❛➂➇➆❫➏☎➃✇➑ ➪✦➈ÑÐ❭➈✍➌❁➈❄➉r➆★➒✇➃❸à✍➈❞➋♠➐✚➉✞➛❤➙➡➞❤➝❤↔ä➧❼➙✥↔❞➙✄✂➡Ï❩➙✥➜r➢✔✂➙❫↔➇↕✖➏❛➋①➃❸➌✚➊➡➒❸➐✚➎❁➈❞➄❛➈✍➚❭➈✍➉r➆★➒ ➒✇➃❸➌✚✃❤➄✍➹❩➏❛➂✚➈✍➌☎➏❛➂✚➈ ✲➍➃❸➏❛➏❛➈✍➉✿➏❛➈✍➌✚➎❁➄✦➏❛➋✠➃❸➌❁➊✍➉❛➈★➆❫➄❛➈✹➎✚➉r➆❫➔☎➆✥➏➍➃❸➊✥➆★➒✇➒❸➘✏➪Ô➃❸➏❛➂☎➏❛➂✚➈ ➌❤➐✚➔✂➶❩➈✍➉✖➋✥➑☛➂❁➋❭➅❁➄✍➹✚➄➍➃❸➌✚➊✍➈❳➏❛➂❁➈❳➅➇➆❫➊r✃♠➈✍➏✦➂❩➆✥➄✖➆❳➄❛➈✍➅❩➆✥➉r➆❫➏❛➈Ò➋♠➅✚➅❩➋♠➉❛➏❛➐✚➌❼➃❸➏❏➘ ➑s➋♠➉❳➐❁➌✚➊✍➋❭➉❛➉❛➈➡➒❷➆❫➏❛➈✍➎✞➷❤➐✚➈✍➐❁➈r➃❸➌❁Ð✏➎❁➈➡➒❷➆✍➘❤➄➮➆❫➏❳➈✥➆❫➊r➂❋➂✚➋♠➅☛➻ ➯➍➌▼➑❏➆✥➊✍➏✍➹✳➃❸➏✂➃❸➄✆➌✚➋♠➏☎➊➡➒❸➈★➆❫➉☎➏❛➂❩➆✥➏✵➏❛➂❤➃❸➄➾➃❸➌✚➊✍➉❛➈✥➆✥➄❛➈✪➃❸➌ ✲➍➃❸➏❛➏❛➈❄➉✆➌✚➈✍➈✍➎ ➋❤➊❄➊✍➐❁➉✍➻❀✑✠➋❫➃❸➌✚Ð✪➏❛➂✚➉❛➋♠➐✚Ð♠➂✞➔✵➋❭➉❛➈✯➂✚➋♠➅✚➄✹➅❁➉❛➋✥➚♠➃❸➎✚➈✍➄✹➔✵➋❭➉❛➈✯➋❭➅❁➅❩➋❭➉❛➏❛➐❁➌❤➃✇➓ ➏➍➃❸➈❄➄✠➑❦➋♠➉❳➄❛➂➇➆❫➉➍➃❸➌✚Ð❁➹❱➆❫➌✚➎Ñ➂✚➈✍➌✚➊✍➈✯➔✲➋♠➉❛➈✂➋♠➅✚➅❩➋♠➉❛➏❛➐✚➌❼➃❸➏➍➃❸➈❄➄✯➑❦➋♠➉❳➉❛➈✍➎✚➐✚➊➡➃❸➌❁Ð ✲❏➃❸➏❛➏❛➈✍➉✍➻❳➁❳➂❁➈✂✃♠➈✍➘☎➃❸➄✠➏❛➋✵➊❄➋♠➉❛➉❛➈➡➒❷➆✥➏❛➈✵➏❛➂❁➈✂➄❛➂➇➆❫➉➍➃❸➌✚Ð✪➈✍➬❤➅❩➈❄➉➍➃❸➈✍➌❁➊❄➈✵➪❳➂❼➃❸➊r➂ ➆✠➅➇➆❫➊r✃♠➈❄➏✳➂➇➆❫➄❳➆✥➏❢➏❛➂❁➈❳➄❛➐✚➊✍➊✍➈✍➄❛➄➍➃❸➚❭➈✹➌❁➋❤➎✚➈✍➄❢➃❸➌✵➃❸➏❛➄✳➅➇➆❫➏❛➂☛➻✿➼Ñ➈❳➊✥➆✍➒✇➒☛➏❛➂❼➃❸➄ ➄❛➊r➂✚➈✍➔✵➈✜✛✻➯✢✛✽✫ ✰ ➻ ✛✻➯■✛✽✫ ✰ ➃❸➄✹➷❤➐➇➆★➒✇➃❸➏r➆❫➏➍➃❸➚❭➈➡➒❸➘▼➚❭➈✍➉❛➘✪➄➍➃❸➔✯➃✇➒❷➆✥➉✯➏❛➋☎➏❛➂✚➈ ✥❸➙r➝♠➥❄↕✻➥ ✥❸➝❤➤r➧✖➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❋➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✵➄❳➑❦➋♠➉✦➔☎➆❫➌❤➐❤➑❏➆❫➊❄➏❛➐❁➉➍➃❸➌✚Ð☎➄❛➘❤➄❛➏❛➈✍➔✵➄ ➎❤➃❸➄❛➊✍➐✚➄❛➄❛➈✍➎❋➃❸➌ ✬Ô➈r➑❦➈✍➉❛➈✍➌✚➊✍➈✆✍✰✂✭✧✭✎❏➻ ➯➍➌ ✛✻➯✢✛✽✫ ✰ ➹❱➪✖➈✵➏❛➉❛➘✞➏❛➋✪➃❸➌❁➎✚➐❁➊❄➈✞✛✻➯✢✛✽✫➴➓❏➄❛➏❏➘♠➒❸➈✆➄❛➂➇➆❫➉➍➃❸➌✚Ð✸➭❦➈✍➷❤➐➇➆★➒ ✲➍➃❸➏❛➏❛➈✍➉☛➑❦➋♠➉❱➆★➒✇➒✚➄❛➋❭➐❁➉❛➊✍➈❄➄✻➃❸➌✯➏❛➂✚➈✖➆✥Ð♠Ð❭➉❛➈✍Ð❤➆✥➏❛➈✖➊➡➒❷➆❫➄❛➄r➺✰➆❫➊✍➉❛➋❭➄❛➄✿➆★➒✇➒✚➏❛➂✚➈✳➂✚➋♠➅✚➄ ➆★➒❸➋❭➌❁Ð✂➏❛➂✚➈➴➅➇➆❫➏❛➂✲➏❛➋✠➔✯➃❸➌❤➃❸➔✯➃❸à✍➈ ✲➍➃❸➏❛➏❛➈❄➉✍➻❢➼Ñ➈❳➎✚➋✹➏❛➂❼➃❸➄✦➆❫➄❢➑❦➋✥➒✇➒❸➋❫➪❳➄✍➻★✛❁➋❭➉ ➈✥➆✥➊➡➂❋➂✚➋♠➅☛➹✻➪✦➈✹➔✵➈✥➆❫➄❛➐✚➉❛➈✵➏❛➂✚➈✵➆✍➚❭➈✍➉r➆❫Ð❭➈✵➎❁➈r➒❷➆★➘❋➄❛➈❄➈✍➌❋➶❤➘❋➅❩➆✥➊➡✃❭➈✍➏❛➄➴➃❸➌ ➈✥➆✥➊➡➂❞➆✥Ð♠Ð❭➉❛➈✍Ð❤➆✥➏❛➈✆➊r➒❷➆❫➄❛➄☎➆✥➏✹➏❛➂❩➆✥➏✯➄❛➪✖➃❸➏❛➊r➂✧➻✏➼Ñ➈☎➏❛➂✚➈✍➌▼➊✍➋❭➔✵➅✚➐❁➏❛➈✵➑❦➋❭➉ ➈✥➆✥➊➡➂✞➅❩➆✥➊➡✃❭➈✍➏✹➏❛➂✚➈✯➎❤➃✇Þ☛➈✍➉❛➈✍➌✚➊✍➈☎➶❩➈✍➏❏➪✖➈✍➈✍➌❋➃❸➏❛➄✠➅➇➆❫➉❛➏➍➃❸➊❄➐❼➒❷➆✥➉✵➎❁➈r➒❷➆★➘✞➆✥➌❁➎ ➏❛➂✚➈✪➊➡➒❷➆✥➄❛➄☎➆✍➚♠➈✍➉r➆✥Ð♠➈♠➻⑧➼Ñ➈✏➆❫➎✚➎❑➭❦➋♠➉✵➄❛➐✚➶✚➏❛➉r➆❫➊✍➏r➺✯➏❛➂❤➃❸➄✵➎❤➃✇Þ☛➈✍➉❛➈✍➌✚➊✍➈❋➏❛➋ ➆✯➽❩➈r➒❸➎ ➃❸➌❋➏❛➂✚➈✯➂✚➈✥➆✥➎❁➈✍➉Ò➋❫➑✦➏❛➂❁➈✂➅➇➆❫➊r✃♠➈✍➏✍➹☛➪❳➂❤➃❸➊➡➂❋➏❛➂❤➐✚➄✹➆✥➊✍➊✍➐✚➔✯➐❤➒❷➆❫➏❛➈✍➄ ➏❛➂✚➈✠➏❛➋♠➏r➆✍➒✰➋❫Þ☛➄❛➈✍➏✿➑❦➋♠➉✖➏❛➂❤➃❸➄❳➅❩➆✥➊➡✃❭➈✍➏✳➑❦➉❛➋❭➔✔➏❛➂✚➈✹➆✍➚♠➈❄➉r➆❫Ð♠➈❳➑❦➋♠➉✿➃❸➏❛➄✖➊➡➒❷➆❫➄❛➄✍➻ ➁❳➂❼➃❸➄✂➽❩➈➡➒❸➎→➆✍➒✇➒❸➋❫➪❳➄☎➈✥➆✥➊➡➂❞➄❛➪✖➃❸➏❛➊➡➂❞➏❛➋❋➊✍➋❭➔✵➅✚➐❁➏❛➈✏➪➴➂✚➈✍➌▼➏❛➂✚➈✪➅➇➆❫➊r✃♠➈❄➏ ➥■✦❩➫❤❮✾✥❸➞✠➂➇➆✍➚♠➈✏➆❫➉❛➉➍➃❸➚❭➈✍➎Ñ➃✇➑✿➃❸➏✹➪✖➈✍➉❛➈✯➃❸➌✚➎✚➈✍➈✍➎▼Ð✥➃❸➚♠➈✍➌❞➆✍➚❭➈✍➉r➆❫Ð❭➈✵➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻ ➁❳➂❁➈✹➄❛➪✖➃❸➏❛➊r➂✪➏❛➂✚➈✍➌✪➄❛➊r➂✚➈✍➎❁➐❤➒❸➈✍➄✹➏❛➂✚➈✹➅❩➆✥➊➡✃❭➈✍➏Ò➆❫➄✖➃✇➑❢➃❸➏Ò➆❫➉❛➉➍➃❸➚❭➈✍➎✞➆❫➏✖➏❛➂❤➃❸➄ ➆✍➚♠➈✍➉r➆✥Ð♠➈✏➈✍➬❤➅❩➈✍➊✍➏❛➈❄➎▼➏➍➃❸➔✵➈✒✰✦➏❛➂❼➃❸➄✯➃❸➄✂➎✚➋♠➌✚➈✪➶❤➘✞➋♠➉❛➎✚➈✍➉➍➃❸➌❁Ð✞➏❛➂✚➈✆➷❤➐✚➈✍➐✚➈ ➶❤➘▼➏❛➂✚➈✍➄❛➈✪➈❄➬❤➅❩➈✍➊✍➏❛➈✍➎ ➆❫➉❛➉➍➃❸➚❭➆★➒➮➏➍➃❸➔✵➈✍➄✵➉r➆✥➏❛➂❁➈❄➉✵➏❛➂❩➆✥➌✸➶❤➘▼➏❛➂✚➈❋➆✥➊✍➏❛➐➇➆★➒ ➆❫➉❛➉➍➃❸➚❭➆✍➒✿➏➍➃❸➔✵➈✍➄✍➻ ➁❢➋➴➏❛➈❄➄❛➏✧➏❛➂❤➃❸➄✿➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✪➹✚➪✖➈❢➂➇➆★➚❭➈✖➄➍➃❸➔✂➐❼➒❷➆✥➏❛➈✍➎✯➃❸➏❛➄✻➅➇➈✍➉➍➑❦➋♠➉❛➔➾➆❫➌✚➊✍➈ ➋♠➌✞➆✵➌✚➈✍➏➍➪✦➋♠➉❛✃❋➆✥➄✠➄❛➂✚➋✥➪➴➌❋➋❭➌✞✛✧➃❸Ð♠➐✚➉❛➈ ✂♠➻❳➁❳➂❤➃❸➄✹➌❁➈❄➏➍➪✖➋❭➉❛✃✪➂➇➆❫➄➴➑s➋♠➐✚➉ ➈✍➷❤➐❤➃❸➚❭➆★➒❸➈✍➌❤➏✫✂✞â✪➶❤➃❸➏✆☎❫➄❛➈✍➊Ñ➃❸➌❤➏❛➈❄➉➍➓➍➄❛➪✖➃❸➏❛➊r➂ã➒✇➃❸➌✚✃❤➄✍➹✂➆✥➌❁➎❑➈✥➆✥➊➡➂✸➒✇➃❸➌✚✃❑➃❸➄ ➄❛➂➇➆❫➉❛➈✍➎✫➶❤➘ ✂✔✕Ñ➱➇➋❫➪❳➄✍➻ ➁❳➂❁➈✍➉❛➈⑧➆❫➉❛➈❭➹✠➃❸➌⑤➏❛➋❭➏r➆★➒❦➹❉✮❏✮Ñ➱❩➋✥➪❳➄☛✰✲➆★➒✇➒✂➋❫➑ ➏❛➂✚➈✍➔✛➂➇➆★➚❭➈✖➏❛➂✚➈➴➄r➆✥➔✵➈✖➄❛➏r➆✥➏➍➃❸➄❛➏➍➃❸➊✥➆★➒☛Ð❭➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌✆➅✚➉❛➋❤➊✍➈✍➄❛➄✖➭❦➎✚➈✍➄❛➊✍➉➍➃❸➶➇➈✍➎ ➃❸➌✸➏❛➂✚➈✆☎✠➅✚➅❩➈✍➌✚➎❼➃❸➬✚➺☎➶❁➐✚➏✵➏❛➂✚➈✍➘▼➏❛➉r➆✍➚♠➈r➒✠➎❤➃✇Þ☛➈✍➉❛➈❄➌❤➏✆➌✚➈✍➏❏➪✖➋♠➉❛✃▼➅➇➆❫➏❛➂✚➄✍➻ ✂✟✮✵➏❛➉r➆✍➚♠➈❄➉❛➄❛➈✵➋♠➌❤➒❸➘❋➋♠➌✚➈✯➃❸➌❤➏❛➈✍➉➍➓❏➄❛➪✖➃❸➏❛➊➡➂❋➒✇➃❸➌✚✃❩➹✯✟☎➏❛➉r➆★➚❭➈✍➉❛➄❛➈✵➏❏➪✖➋✵➃❸➌❤➏❛➈✍➉➍➓ ➄❛➪✖➃❸➏❛➊➡➂✞➒✇➃❸➌❁✃❤➄✍➹❋✟❋➏❛➉r➆✍➚♠➈✍➉❛➄❛➈✲➏❛➂❁➉❛➈✍➈✵➃❸➌❤➏❛➈✍➉➍➓❏➄❛➪✖➃❸➏❛➊r➂①➒✇➃❸➌✚✃❤➄✍➹✦➆❫➌✚➎✁✮✆➏❛➉r➆✍➓ ➚♠➈❄➉❛➄❛➈☎➆★➒✇➒❢➑❦➋❭➐❁➉❳➃❸➌❤➏❛➈✍➉➍➓❏➄❛➪✖➃❸➏❛➊➡➂Ñ➒✇➃❸➌✚✃❤➄✍➻✂➁✿➆❫➶❤➒❸➈ ✮✵➎❼➃❸➄❛➅❤➒❷➆✍➘❤➄✯➏❛➂✚➈✵➔✵➈★➆❫➌ ➆❫➌✚➎✚✓✖✓❤➻✼✓✻✳ ➏❛➂❞➅❩➈✍➉❛➊❄➈✍➌❤➏➍➃✇➒❸➈Ñ➷❤➐✚➈✍➐✚➈➡➃❸➌✚Ð▼➎❁➈r➒❷➆★➘❤➄✵➑❦➋❭➉☎➆✪➄➍➃❸➌✚Ð❫➒❸➈✞➄r➆❫➔✵➅❤➒❸➈ ➱➇➋❫➪✸➑s➋♠➉✖➈✥➆✥➊➡➂✏➅❩➆✥➏❛➂✆➒❸➈✍➌✚Ð♠➏❛➂✞➭❦➏❛➂❁➈✹➎➇➆❫➏r➆✹➑❦➉❛➋♠➔ ➏❛➂❁➈Ò➋♠➏❛➂✚➈✍➉✖➱❩➋✥➪❳➄❳➆❫➉❛➈ ➄➍➃❸➔✯➃✇➒❷➆✥➉r➺➡➻▼➼✞➈✆➊❄➋♠➔✵➅➇➆❫➉❛➈✏➏❛➂❁➈☎➼✛✂✯➹❋✛✻➯✢✛✦✫✯➹❢➆✥➌❁➎✚✛✻➯■✛✽✫ ✰ ➆★➒❸Ð❭➋❫➓ ➉➍➃❸➏❛➂✚➔✵➄✆➭❦➪❳➂✚➈✍➉❛➈✵➪✖➈✂➂➇➆★➚❭➈☎➐❁➄❛➈✍➎✞➈✍➷❤➐➇➆★➒✦➊➡➒❸➋❤➊r✃Ñ➉r➆✥➏❛➈✍➄✠➃❸➌▼➏❛➂✚➈✵➼✛✂ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔☎➺➡➻❞➸❳➋♠➏❛➈☎➏❛➂➇➆❫➏✂➏❛➂❁➈☎➔✵➈✥➆✥➌❞➎❁➈➡➒❷➆✍➘❤➄☎➆✥➉❛➈☎➊✍➋♠➔✵➅➇➆❫➉r➆✥➶❼➒❸➈✪➃❸➌ ➆★➒✇➒✿➏❛➂✚➉❛➈✍➈✹➊✥➆✥➄❛➈✍➄✍➻❳➼✸➂❤➃✇➒❸➈✵➏❛➂✚➈✠✓✬✓❼➻ ✓ ✳ ➏❛➂✪➅❩➈❄➉❛➊✍➈✍➌❤➏➍➃✇➒❸➈☎➎❁➈r➒❷➆★➘❤➄✠➃❸➌✚➊✍➉❛➈✥➆✥➄❛➈ ➪✖➃❸➏❛➂☎➅❩➆✥➏❛➂➾➒❸➈❄➌❁Ð❭➏❛➂✆➑❦➋❭➉✖➆✍➒✇➒✻➏❛➂✚➉❛➈✍➈➮➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✲➄✍➹☛➏❛➂❁➈❳➉r➆❫➏❛➈Ò➋❫➑✰Ð♠➉❛➋✥➪❳➏❛➂ ➃❸➄✠➔✂➐✚➊➡➂✏➄❛➔☎➆★➒✇➒❸➈✍➉✹➪✖➃❸➏❛➂✏➏❛➂❁➈✵✛✻➯■✛✽✫ ✰ ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✪➻ ☎✠➄✹➏❛➂✚➈✆➄➍➃❸➔✂➐❤➒❷➆❫➏➍➃❸➋♠➌→➄❛➂❁➋✥➪❳➄✍➹✿➏❛➂❁➈✲➈➡Þ☛➈✍➊✍➏✂➋❫➑✄✛✻➯✢✛✽✫ ✰ ➹ ➆❫➄✹➊✍➋❭➔✯➓ ➅➇➆❫➉❛➈✍➎✵➏❛➋❀✛✻➯✢✛✦✫✯➹✍➃❸➄✳➏❛➋Ò➄➍➒✇➃❸Ð♠➂❤➏➍➒❸➘✯➃❸➌✚➊✍➉❛➈✥➆✥➄❛➈✠➏❛➂✚➈✖➔✵➈✥➆✥➌✯➎✚➈➡➒❷➆✍➘☎➆❫➌✚➎✾✲➍➃❸➏➍➓ ➏❛➈✍➉✖➋✥➑☛➱❩➋✥➪❳➄✖➋♠➌☎➄❛➂✚➋♠➉❛➏✖➅➇➆❫➏❛➂✚➄✍➹❩➄➍➒✇➃❸Ð❭➂❤➏➍➒❸➘✪➎✚➈✍➊✍➉❛➈✥➆✥➄❛➈✹➏❛➂❁➈Ò➔✵➈✥➆✥➌✆➎✚➈➡➒❷➆✍➘ ➆❫➌✚➎✞➄➍➃❸Ð♠➌❤➃✇➽❩➊✥➆✥➌❤➏➍➒③➘❞➎✚➈✍➊✍➉❛➈✥➆✥➄❛➈✵➏❛➂✚➈❃✲➍➃❸➏❛➏❛➈✍➉✹➋✥➑✳➱➇➋✥➪➴➄✹➋❭➌ ➒❸➋❭➌❁Ð❋➅➇➆❫➏❛➂✚➄✍➹ ➪❳➂❼➃❸➊r➂❞➔✵➈✥➆✥➌❁➄✵➏❛➂➇➆❫➏✲➏❛➂❁➈☎➋✥➚♠➈✍➉r➆✍➒✇➒Ò➎❁➈➡➒❷➆✍➘❞➶❩➋❭➐❁➌✚➎✸Ð❭➋❤➈✍➄✵➎✚➋❫➪❳➌❞➆✥➌❁➎ ➏❛➂✚➈✠➅✚➉❛➈✍➊➡➃❸➄➍➃❸➋♠➌❋➋✥➑✰➈✍➄❛➏➍➃❸➔☎➆❫➏➍➃❸➋❭➌✪Ð❭➋❤➈✍➄✖➐✚➅✆➋❭➌✵➒❸➋♠➌✚Ð✵➅❩➆✥➏❛➂❁➄✍➻✿➼→➂❁➈✍➌✏➪✖➈ ➊✍➋❭➔✵➅❩➆✥➉❛➈✖➏❛➂✚➈✳➃❸➔✵➅❤➒❸➈✍➔✵➈✍➌❤➏r➆✥➏➍➃❸➋♠➌✪➋✥➑☛➏❛➂✚➈✖➏➍➪✖➋❳➄❛➊r➂❁➈❄➔✵➈✍➄✍➹❤➏❛➂✚➈✍➘✹➎❤➃✇Þ☛➈✍➉❛➄ ➃❸➌✪➋❭➌❁➈✠➃❸➔✲➅❩➋♠➉❛➏r➆✥➌❤➏✠➪❳➆✍➘✆☎✲➏❛➂❁➈✹➷❤➐✚➈✍➐✚➈✹➔☎➆❫➌➇➆❫Ð❭➈✍➔✵➈✍➌❤➏❳➎❤➃❸➄❛➊➡➃❸➅❤➒✇➃③➌✚➈✵➃❸➄
no longer trivial (add the packet to the tail of the queue for t agram traffic to the lowest priority class. There are K other priority levels above the by deadline, where the deadline is explicitly computed by At the service interface, we provide K wid taking the actual arrival time, adjusting this by the offset in get delay bounds Di for predicted service (at a particular he packet header to find the expected arrival time, and the using this to order the bility of the diferent k classes. These n to separate the traffic for ore expensive processing overhead, but we believe that ef- of the actual delivered delays. Rather, they are a priori up ficient coding methods can implement this in software with per bounds and the network tries, through admission cont rol the same performance as current packet switches achieve. polici keep queueing delays at each switch for a part We have now extended our predicted service class lar class i well below these bounds di, We ment ioned earlier multiple hops, using FIFO+ as an explicit means to min- that adaptive applications have limits to their adapta bility imize the jitter by obtaining as much benefit as possible these bounds Di are indicative of such limits. A predicted from sharing. Recall that for guaranteed service, the servic service flow is assigned a priority level at each switch(not specified by the worst-case bounds and thus the focus is necessarily the same level in every switch); the a priori de scheduling algorithms that provide isolation between the lay bound advertised to a predicted service flow is the sum various flows. In contrast, for predicted service we assume of the appropriate Di along the path. The delay bound ad that (1)adequate isolation is being provided by the enfo vertised to a guaranteed service flow is the Parekh-Galla ent of traffic filters before or at the ent rance to the net work bound ther than by the scheduling algorithm itself, and (2)net- This scheme has the pro blem that, since delay is additive rather than being wors asking for a particular Di at a given swit ch does not directly similar to those of the recent past. For predicted service, mean that D is the t arget delay bound for the path as the challenge is to share the link effectively in a way that whole. Rather, it is necessary to add up the target delays the play-back point. As we have seen, FIFO is at each hop to find the target upper bound for the pa th sharing mechanism. The modification of FIFO We expect the true post facto bounds over a long path to be merely extends the concept of sh aring from sharing between significant ly lower than the sum of the bounds D at flows at a single hop to sharing bet ween hops hop. But we suggest that, since this is an adaptive th empt to ch trol 7 Unified Schedu ling algorit the service to great precisio d thus should just use the sum of the Dis as the advertised bound. have presented schedulin Consider in more detail how the priority scheme works algorithms that each handle a single kind of service com If the highest priority class has a moment ary need for extra mit ment. In this section we combine these algorithms into a band width due to a burst b ral of unified scheduling algorithm that handles guaranteed, pre- the bandwidth from the lower cl es. The next class tht dicted, and datagram a base line of operation th ate jitter of the Consider a set of real-time flows, some requesting guaran- higher class. This gets factored together with the aggregate teed service and some requesting predicted service, and also how this fits into our overall service architecture en disc s a set of datagram sources. We first describe the scheduling second class. This cascades down to the datagram traffic orithm as im plemented at each switch and the rhich gets whatever bandwidth is leftover and suffers from the accumulated jitter. As we argue later, the datagram The sched straight- traffic should probably be given an average rate of at least forward. The basic idea is that we must isolate each of the both to insure that it makes some progress on tI uaranteed service flows from each other and from the pre average and to provide a reasonable pool of bandwidth fo dicted service flows(and the datagram traffic). Therefore the higher priority traffic to borrow from during momental we use the time-stamp based WFQ scheme as a frame wor overloads into which we fit the other scheduling algorithms.Each For a lower priority class, what source of jitter will dom arant ed service client a has a separate wFQ fow with late its observed behavior: its intrinsic aggregate behavio some clock rate r. All of the predicted service and data- or the jitter shifted from the higher priority classes? if the gram service traffic is assigned to a pseudo WFQ flow, call target goals for jitter are widely spaced (and for the pr it flow 0, with, at each link, ro=u- ra where the pose of rough estimation as we suggested above they prol that link. Inside this flow 0. there are a number of strict exported jitter from the higher priority class should be ar ity classes, and wit hin each priority class we operat order of magnit ude less than the intrinsic behavior of tI the FIFo+ algorithm. Once class, and the classes should usually operate more or less tive service flow (and also the independently. Thus, a particular class is isolated from t le vel at each switch lower priority classes by the priority scheduling algorith discuss how this algorithm fits into our and is in effect isolated from the higher priority classes be We have discussed the funct ion of the FiFo+ scheme particular class above. What is the role of the priority cl g algorithm us- above that the effect of priority is to shift the jitter of higher ing the same simulation configuration as used for Table 2. ty classes. We assign that has 22 reaLtime flows with identical statistical gener- tor but which tr different network path tuted fer wegernags otner iso ating mechanisms t iese realisme lats we连过2amcm guaranteed service clients; 3 of these have a clock rate equal
➌✚➋✯➒❸➋❭➌❁Ð❭➈✍➉❳➏❛➉➍➃❸➚♠➃❷➆★➒❳➭❏➆❫➎✚➎☎➏❛➂❁➈✹➅➇➆❫➊r✃♠➈❄➏✖➏❛➋✯➏❛➂✚➈✹➏r➆★➃✇➒✰➋✥➑❢➏❛➂✚➈✠➷❤➐✚➈✍➐❁➈✠➑s➋♠➉ ➏❛➂✚➈❋➊➡➒❷➆✥➄❛➄r➺➾➶✚➐✚➏➾➃❸➌✚➄❛➏❛➈✥➆✥➎❑➉❛➈✍➷❤➐❼➃❸➉❛➈❄➄✪➏❛➂➇➆❫➏☎➏❛➂❁➈✏➷❤➐❁➈❄➐❁➈✞➶❩➈✪➋❭➉❛➎❁➈❄➉❛➈✍➎ ➶❤➘⑧➎❁➈✥➆✥➎❼➒✇➃❸➌✚➈♠➹Ò➪❳➂❁➈✍➉❛➈☎➏❛➂✚➈✪➎✚➈✥➆❫➎❤➒✇➃❸➌✚➈ ➃❸➄✵➈❄➬❤➅❼➒✇➃❸➊➡➃❸➏➍➒❷➘✸➊❄➋♠➔✵➅✚➐❁➏❛➈❄➎✸➶❤➘ ➏r➆✥✃♠➃❸➌❁Ð✲➏❛➂❁➈Ò➆❫➊❄➏❛➐❩➆✍➒✰➆❫➉❛➉➍➃❸➚❭➆★➒✰➏➍➃❸➔✵➈❭➹❩➆✥➎✣✲➳➐❁➄❛➏➍➃❸➌✚Ð✵➏❛➂❤➃❸➄✖➶❤➘✯➏❛➂✚➈✠➋✥Þ☛➄❛➈✍➏❢➃❸➌ ➏❛➂✚➈❳➅❩➆✥➊➡✃❭➈✍➏✖➂✚➈✥➆❫➎✚➈✍➉✖➏❛➋✠➽➇➌✚➎✵➏❛➂❁➈✲➙✄✂❄Ï➇➙r➤✥↕◗➙r➞✿➆❫➉❛➉➍➃❸➚❭➆★➒✰➏➍➃❸➔✵➈❭➹✚➆❫➌✚➎☎➏❛➂❁➈✍➌ ➐✚➄➍➃❸➌✚Ð❋➏❛➂❼➃❸➄✹➏❛➋☎➋♠➉❛➎✚➈✍➉Ò➏❛➂❁➈✵➷❤➐✚➈✍➐✚➈♠➻✹➁❳➂❤➃❸➄✹➂❩➆✥➄✠➏❛➂✚➈✯➅❩➋❭➄❛➄➍➃❸➶❼➃✇➒❸➃❸➏➍➘⑧➋❫➑Ô➆ ➔✲➋♠➉❛➈✠➈❄➬❤➅❩➈✍➌✚➄➍➃❸➚♠➈✂➅❁➉❛➋❤➊✍➈❄➄❛➄➍➃❸➌❁Ð☎➋✥➚♠➈✍➉❛➂✚➈✥➆✥➎✧➹➇➶❁➐✚➏✖➪✖➈Ò➶❩➈➡➒✇➃❸➈✍➚❭➈✯➏❛➂➇➆❫➏✖➈r➑③➓ ➽➇➊➡➃❸➈✍➌❤➏Ò➊✍➋❤➎❤➃❸➌❁Ð✏➔✵➈✍➏❛➂✚➋❤➎❁➄Ò➊✥➆❫➌☎➃❸➔✵➅❼➒❸➈❄➔✵➈✍➌❤➏✹➏❛➂❤➃❸➄➴➃❸➌❋➄❛➋❫➑s➏➍➪❳➆✥➉❛➈✹➪Ô➃❸➏❛➂ ➏❛➂✚➈✹➄r➆❫➔✲➈✯➅❩➈❄➉➍➑❦➋♠➉❛➔☎➆✥➌❁➊❄➈✂➆❫➄✠➊❄➐❁➉❛➉❛➈✍➌❤➏❳➅➇➆❫➊r✃♠➈✍➏Ò➄❛➪✖➃❸➏❛➊➡➂✚➈✍➄Ò➆❫➊r➂❼➃❸➈❄➚♠➈♠➻ ➼✞➈▼➂❩➆✍➚❭➈Ý➌✚➋✥➪❐➈✍➬❤➏❛➈✍➌✚➎❁➈✍➎✷➋♠➐✚➉❋➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✽➄❛➈✍➉❛➚♠➃❸➊❄➈Ý➊r➒❷➆❫➄❛➄✞➏❛➋ ➔✂➐❤➒❸➏➍➃❸➅❼➒❸➈✞➂❁➋❭➅❁➄✍➹✦➐❁➄➍➃❸➌✚Ð✫✛✻➯■✛✽✫ ✰ ➆❫➄✲➆❫➌❞➈✍➬❤➅❤➒✇➃❸➊➡➃❸➏✪➔✵➈★➆❫➌✚➄✵➏❛➋Ñ➔✂➃❸➌❼➓ ➃❸➔✂➃❸à✍➈▼➏❛➂✚➈ ✲➍➃❸➏❛➏❛➈❄➉❋➶❤➘✸➋♠➶✚➏r➆★➃❸➌❤➃❸➌✚Ð✫➆❫➄✪➔✂➐✚➊➡➂✫➶❩➈✍➌✚➈➡➽➇➏✞➆❫➄✪➅➇➋♠➄❛➄➍➃❸➶❤➒❸➈ ➑s➉❛➋♠➔✽➄❛➂❩➆✥➉➍➃❸➌❁Ð✚➻✾✬✖➈✍➊✥➆✍➒✇➒✰➏❛➂❩➆✥➏ ➑❦➋❭➉✳Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎✪➄❛➈✍➉❛➚♠➃❸➊✍➈❭➹❁➏❛➂✚➈❳➄❛➈✍➉❛➚♠➃❸➊✍➈ ➃❸➄✹➄❛➅➇➈✍➊➡➃✇➽➇➈✍➎❞➶❤➘❋➏❛➂❁➈✂➪✖➋♠➉❛➄❛➏➍➓❏➊✥➆✥➄❛➈✵➶❩➋❭➐❁➌✚➎❁➄✲➆❫➌✚➎✞➏❛➂❤➐❁➄✹➏❛➂✚➈✯➑❦➋❤➊✍➐✚➄✠➃❸➄ ➋❭➌☎➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð✏➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✲➄✠➏❛➂➇➆❫➏✖➅✚➉❛➋✥➚♠➃❸➎❁➈☎➢ ➥❄➫✖✥❸➝❤↕❦➢③➫❤↔✖➶❩➈✍➏❏➪✖➈✍➈✍➌☎➏❛➂✚➈ ➚❭➆✥➉➍➃❸➋♠➐✚➄P➱➇➋✥➪➴➄✍➻✲➯➍➌✞➊✍➋♠➌❤➏❛➉r➆✥➄❛➏✍➹✧➑s➋♠➉✹➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎▼➄❛➈✍➉❛➚♠➃❸➊✍➈☎➪✖➈➾➆❫➄❛➄❛➐✚➔✵➈ ➏❛➂➇➆❫➏Ò➭✢✂✥➺✿➆❫➎✚➈✍➷❤➐❩➆✥➏❛➈➴➃❸➄❛➋✥➒❷➆❫➏➍➃❸➋♠➌✪➃❸➄✦➶❩➈➡➃❸➌✚Ð✯➅✚➉❛➋✥➚♠➃❸➎❁➈✍➎✏➶❤➘✵➏❛➂✚➈➴➈❄➌❼➑❦➋❭➉❛➊✍➈➡➓ ➔✲➈✍➌❤➏❢➋❫➑➇➏❛➉r➆★Ó☎➊❢➽✚➒❸➏❛➈✍➉❛➄✳➶➇➈➡➑❦➋♠➉❛➈✿➋♠➉✿➆✥➏✻➏❛➂✚➈✖➈✍➌❤➏❛➉r➆✥➌❁➊✍➈✦➏❛➋➴➏❛➂✚➈✖➌✚➈✍➏➍➪✦➋♠➉❛✃ ➉r➆✥➏❛➂❁➈❄➉✠➏❛➂➇➆❫➌❋➶❤➘✪➏❛➂✚➈✹➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✞➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔❅➃❸➏❛➄❛➈➡➒✇➑❛➹✰➆✥➌❁➎▼➭ ✮❭➺Ô➌✚➈✍➏➍➓ ➪✦➋♠➉❛✃❞➊✍➋♠➌✚➎❤➃❸➏➍➃❸➋♠➌✚➄✍➹✯➉r➆✥➏❛➂❁➈❄➉✆➏❛➂➇➆❫➌→➶❩➈➡➃❸➌✚Ð✸➪✖➋❭➉❛➄❛➏➍➓❏➊✥➆❫➄❛➈♠➹Ò➆✥➉❛➈❋➉r➆❫➏❛➂✚➈✍➉ ➄➍➃❸➔✂➃✇➒❷➆❫➉✪➏❛➋❋➏❛➂✚➋♠➄❛➈❋➋✥➑Ò➏❛➂❁➈✏➉❛➈✍➊✍➈✍➌❤➏✵➅➇➆❫➄❛➏✍➻ ✛❁➋♠➉✂➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎✸➄❛➈✍➉❛➚♠➃❸➊❄➈♠➹ ➏❛➂✚➈☎➊➡➂➇➆★➒✇➒❸➈❄➌❁Ð❭➈ ➃❸➄✂➏❛➋▼➥✢✦❩➝❤➜❏➙✹➏❛➂❁➈✵➒✇➃❸➌✚✃Ý➈rÞ☛➈✍➊✍➏➍➃❸➚❭➈➡➒❸➘▼➃❸➌✸➆✪➪Ô➆★➘✞➏❛➂➇➆❫➏ ✔✂➢③↔❩➢✔✯➢☎✄✥➙✍➥✠➏❛➂✚➈☎➅❼➒❷➆✍➘♠➓❏➶➇➆❫➊r✃▼➅❩➋❫➃❸➌❤➏✍➻✤☎✠➄✹➪✖➈☎➂➇➆✍➚♠➈☎➄❛➈✍➈✍➌☛➹★✛✻➯✢✛✽✫✔➃❸➄ ➆✂➈➡Þ☛➈✍➊✍➏➍➃❸➚❭➈✵➄❛➂➇➆❫➉➍➃❸➌✚Ð✪➔✵➈✍➊r➂➇➆❫➌❤➃❸➄❛➔✪➻❳➁❳➂❁➈✂➔✵➋❤➎❤➃✇➽❩➊✥➆✥➏➍➃❸➋♠➌▼➋✥➑★✛✻➯✢✛✦✫ ✰ ➔✲➈✍➉❛➈➡➒❸➘✵➈✍➬❤➏❛➈❄➌❁➎✚➄✖➏❛➂✚➈✠➊✍➋❭➌❁➊✍➈❄➅❁➏✖➋✥➑☛➄❛➂➇➆❫➉➍➃❸➌✚Ð✯➑❦➉❛➋♠➔✽➄❛➂➇➆❫➉➍➃❸➌✚Ð✵➶❩➈✍➏❏➪✖➈✍➈✍➌ ➱➇➋✥➪❳➄✠➆✥➏Ò➆✯➄➍➃❸➌❁Ð✥➒❸➈✵➂✚➋♠➅❋➏❛➋✵➄❛➂➇➆❫➉➍➃❸➌✚Ð☎➶❩➈✍➏➍➪✦➈✍➈✍➌❋➂✚➋♠➅✚➄✍➻ ✁✖Ø❤Ü✂❍✺♠Ú ✏❙➀✬❚●✺♠Ú➇Û✻❂ Ü❸Ø✆☎✏❹❏❂ ☎✚Ù✚❾◗Ü✇❽ ❚❃❆ ➯❏➌✪➏❛➂✚➈✠➅✚➉❛➈✍➚♠➃❸➋♠➐✚➄✹➏❛➂✚➉❛➈✍➈✹➄❛➈✍➊✍➏➍➃❸➋❭➌❁➄Ò➪✖➈✠➂➇➆✍➚♠➈✹➅✚➉❛➈✍➄❛➈✍➌❤➏❛➈✍➎✏➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð ➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✵➄✏➏❛➂❩➆✥➏☎➈✥➆❫➊r➂❞➂❩➆✥➌❁➎❤➒❸➈▼➆✞➄➍➃❸➌❁Ð✥➒❸➈Ñ✃♠➃❸➌✚➎✸➋✥➑✹➄❛➈✍➉❛➚♠➃❸➊✍➈❋➊❄➋♠➔✯➓ ➔✂➃❸➏❛➔✵➈✍➌❤➏✍➻✿➯➍➌✵➏❛➂❤➃❸➄✖➄❛➈✍➊❄➏➍➃❸➋♠➌☎➪✖➈❳➊✍➋♠➔✂➶❤➃❸➌❁➈Ò➏❛➂❁➈✍➄❛➈Ò➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✵➄✖➃❸➌❤➏❛➋✲➆ ➐✚➌❤➃✇➽❩➈✍➎❞➄❛➊r➂❁➈❄➎❁➐❤➒✇➃❸➌❁Ð①➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➏❛➂❩➆✥➏✹➂❩➆✥➌❁➎❤➒❸➈✍➄✵Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎✧➹❢➅✚➉❛➈➡➓ ➎❤➃❸➊✍➏❛➈✍➎☛➹✰➆❫➌✚➎❋➎➇➆❫➏r➆❫Ð❭➉r➆❫➔ ➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻ ✖➋♠➌✚➄➍➃❸➎❁➈❄➉✿➆Ô➄❛➈❄➏✻➋❫➑✚➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✳➱➇➋✥➪➴➄✍➹❭➄❛➋♠➔✲➈✳➉❛➈✍➷❤➐✚➈✍➄❛➏➍➃❸➌✚Ð✠Ð❭➐❩➆✥➉r➆❫➌❤➓ ➏❛➈❄➈✍➎☎➄❛➈✍➉❛➚♠➃❸➊✍➈✂➆❫➌✚➎☎➄❛➋♠➔✵➈Ò➉❛➈✍➷❤➐✚➈✍➄❛➏➍➃❸➌❁Ð✲➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✏➄❛➈✍➉❛➚♠➃❸➊✍➈♠➹❙➆❫➌✚➎✏➆★➒❸➄❛➋ ➆✵➄❛➈✍➏❳➋✥➑✿➎❩➆✥➏r➆❫Ð♠➉r➆✥➔❅➄❛➋♠➐✚➉❛➊✍➈✍➄✍➻❳➼✞➈✹➽❩➉❛➄❛➏Ò➎❁➈✍➄❛➊❄➉➍➃❸➶❩➈✵➏❛➂✚➈✹➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð ➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➆❫➄✹➃❸➔✵➅❼➒❸➈❄➔✵➈✍➌❤➏❛➈✍➎→➆❫➏✹➈★➆❫➊r➂Ñ➄❛➪✖➃❸➏❛➊r➂❞➆❫➌✚➎▼➏❛➂✚➈✍➌▼➎❤➃❸➄❛➊✍➐❁➄❛➄ ➂✚➋✥➪⑤➏❛➂❤➃❸➄➴➽➇➏❛➄✖➃❸➌❤➏❛➋☎➋♠➐✚➉➴➋✥➚❭➈✍➉r➆★➒✇➒✿➄❛➈✍➉❛➚♠➃❸➊❄➈✵➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈♠➻ ➁❳➂❁➈✿➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✂➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔ ➆❫➏❱➆Ô➄➍➃❸➌✚Ð❫➒❸➈✦➄❛➪Ô➃❸➏❛➊r➂✹➃❸➄✻➷❤➐❤➃❸➏❛➈✖➄❛➏❛➉r➆★➃❸Ð❭➂❤➏➍➓ ➑s➋♠➉❛➪❳➆✥➉❛➎✧➻❳➁❳➂✚➈✹➶❩➆✥➄➍➃❸➊✯➃❸➎❁➈★➆➾➃❸➄✠➏❛➂➇➆❫➏Ò➪✖➈✹➔✂➐❁➄❛➏❳➃❸➄❛➋❫➒❷➆❫➏❛➈✲➈✥➆❫➊r➂❋➋❫➑✿➏❛➂✚➈ Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎Ñ➄❛➈✍➉❛➚♠➃❸➊❄➈P➱➇➋✥➪➴➄❳➑❦➉❛➋♠➔✾➈★➆❫➊r➂❋➋♠➏❛➂✚➈✍➉✹➆✥➌❁➎✪➑❦➉❛➋❭➔✾➏❛➂❁➈✹➅✚➉❛➈➡➓ ➎❤➃❸➊✍➏❛➈✍➎→➄❛➈✍➉❛➚♠➃❸➊✍➈☎➱❩➋✥➪❳➄✪➭❏➆❫➌✚➎❞➏❛➂❁➈☎➎➇➆❫➏r➆✥Ð♠➉r➆❫➔ ➏❛➉r➆★Ó☎➊★➺➡➻Ñ➁❳➂✚➈✍➉❛➈➡➑❦➋❭➉❛➈ ➪✦➈✹➐❁➄❛➈✂➏❛➂❁➈✂➏➍➃❸➔✵➈➡➓❏➄❛➏r➆✥➔✵➅❋➶❩➆✥➄❛➈✍➎✞➼✛✂✛➄❛➊r➂❁➈❄➔✵➈✵➆✥➄✹➆P➑s➉r➆❫➔✵➈✍➪✦➋♠➉❛✃ ➃❸➌❤➏❛➋⑧➪➴➂❤➃❸➊r➂❑➪✖➈✆➽➇➏✵➏❛➂✚➈❋➋♠➏❛➂✚➈✍➉☎➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❑➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✵➄☎✄❁➻✛✎✳➆✥➊➡➂ Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎Ý➄❛➈❄➉❛➚♠➃❸➊✍➈☎➊➡➒✇➃❸➈✍➌❤➏✵✻❞➂❩➆✥➄✵➆☎➄❛➈✍➅❩➆✥➉r➆❫➏❛➈☎➼✛✂ ➱➇➋✥➪✽➪Ô➃❸➏❛➂ ➄❛➋❭➔✵➈☎➊➡➒❸➋❤➊r✃Ñ➉r➆✥➏❛➈✵✠ ✹ ➻✜☎➴➒✇➒✦➋❫➑❳➏❛➂✚➈☎➅❁➉❛➈❄➎❼➃❸➊✍➏❛➈❄➎Ý➄❛➈❄➉❛➚♠➃❸➊✍➈✏➆❫➌✚➎▼➎❩➆✥➏r➆★➓ Ð❭➉r➆❫➔ ➄❛➈✍➉❛➚♠➃❸➊✍➈✵➏❛➉r➆✍Ó☎➊✹➃❸➄✂➆❫➄❛➄➍➃❸Ð♠➌✚➈✍➎▼➏❛➋✏➆✵➅✚➄❛➈✍➐❁➎✚➋✪➼✛✂✛➱❩➋✥➪✹➹☛➊✥➆★➒✇➒ ➃❸➏✵➱➇➋✥➪ ✕❤➹❳➪Ô➃❸➏❛➂☛➹Ò➆❫➏☎➈✥➆✥➊➡➂❞➒✇➃❸➌❁✃❩➹ ✠ ✣ ✥✧✽ ✲ ✞ ✹ ✠ ✹ ➪➴➂✚➈✍➉❛➈❋➏❛➂✚➈ ➄❛➐✚➔ ➃❸➄❳➋✥➚♠➈❄➉❳➆★➒✇➒✰➏❛➂✚➈✠Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎✏➄❛➈✍➉❛➚♠➃❸➊✍➈✠➱➇➋❫➪❳➄✖➅➇➆❫➄❛➄➍➃❸➌✚Ð☎➏❛➂❁➉❛➋❭➐❁Ð❭➂ ➏❛➂➇➆❫➏✂➒✇➃❸➌❁✃❩➻❞➯❏➌❁➄➍➃❸➎✚➈❋➏❛➂❼➃❸➄✂➱❩➋✥➪◗✕❼➹✿➏❛➂✚➈✍➉❛➈❋➆✥➉❛➈✏➆ ➌❤➐✚➔✯➶➇➈✍➉✵➋❫➑➮➄❛➏❛➉➍➃❸➊✍➏ ➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘✸➊➡➒❷➆✥➄❛➄❛➈✍➄✍➹Ò➆❫➌✚➎✸➪✖➃❸➏❛➂❤➃❸➌✸➈✥➆✥➊➡➂→➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘→➊➡➒❷➆❫➄❛➄☎➪✖➈✏➋♠➅❩➈✍➉r➆✥➏❛➈ ➏❛➂✚➈✜✛✻➯✢✛✽✫ ✰ ➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔✪➻ ✫✠➌✚➊✍➈✯➪✦➈✹➂❩➆✍➚♠➈✲➆❫➄❛➄➍➃❸Ð❭➌❁➈❄➎▼➈✥➆❫➊r➂❋➅✚➉❛➈✍➎❼➃❸➊r➓ ➏➍➃❸➚❭➈✠➄❛➈✍➉❛➚♠➃❸➊❄➈➴➱➇➋❫➪✫➭❏➆❫➌✚➎✏➆★➒❸➄❛➋✂➏❛➂❁➈❳➎➇➆❫➏r➆❫Ð❭➉r➆❫➔ ➏❛➉r➆★Ó☎➊✥➺✻➏❛➋✂➆✠➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘ ➒❸➈❄➚♠➈➡➒✦➆✥➏✖➈✥➆❫➊r➂❋➄❛➪✖➃❸➏❛➊➡➂☛➹❩➏❛➂✚➈✹➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✞➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔ ➃❸➄❳➊✍➋❭➔✵➅❼➒❸➈❄➏❛➈➡➒❸➘ ➎✚➈➡➽➇➌❁➈❄➎✧➻❑➼✞➈❋➌✚➋❫➪✾➎❤➃❸➄❛➊✍➐✚➄❛➄✪➂✚➋✥➪✾➏❛➂❼➃❸➄✏➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➽➇➏❛➄✵➃❸➌❤➏❛➋▼➋♠➐✚➉ ➋✥➚❭➈✍➉r➆★➒✇➒✿➄❛➈✍➉❛➚♠➃❸➊❄➈✵➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈♠➻ ➼✞➈❋➂❩➆✍➚♠➈❋➎❤➃❸➄❛➊✍➐✚➄❛➄❛➈✍➎✫➏❛➂❁➈✪➑❦➐✚➌✚➊✍➏➍➃❸➋♠➌❑➋✥➑✂➏❛➂❁➈✞✛✻➯■✛✽✫ ✰ ➄❛➊r➂❁➈❄➔✵➈ ➆✥➶❩➋✥➚♠➈♠➻➮➼✸➂➇➆❫➏➴➃❸➄✹➏❛➂✚➈✂➉❛➋❫➒❸➈✵➋✥➑✿➏❛➂❁➈✂➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✞➊➡➒❷➆❫➄❛➄❛➈❄➄✄✵✬Ô➈❄➔✵➈✍➔✂➶❩➈✍➉ ➆✥➶❩➋✥➚♠➈❳➏❛➂➇➆❫➏✿➏❛➂❁➈❳➈➡Þ☛➈❄➊✍➏✳➋✥➑☛➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘✂➃❸➄✖➏❛➋✠➄❛➂❤➃✇➑❦➏✿➏❛➂❁➈✼✲➍➃❸➏❛➏❛➈✍➉✿➋❫➑❙➂❼➃❸Ð❭➂❁➈✍➉ ➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘❋➊➡➒❷➆❫➄❛➄Ò➏❛➉r➆★Ó☎➊✹➏❛➋✲➏❛➂❁➈✠➒❸➋✥➪✖➈✍➉Ò➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✏➊➡➒❷➆❫➄❛➄❛➈✍➄✍➻❳➼✞➈✂➆❫➄❛➄➍➃❸Ð❭➌ ✄ ✰✹ç☛è❦ì✥ô✍ò✟✩ õ✖ÿ✥ô✍í❦ç☛í❦ì❭é➡í❁í❦ì✥ç❛ê❦ç✰é➡ê❦ç❩ó♠ç❛ê❦ì❭é➡ó✥è❁ô❄í◗ì✥ç❛ê ✘ è❦ô☛✩ é➡í✆✘ ÿ✥û✼✚✖ç❛ë➍ì❭é➡ÿ✟✘ è▲✚✦è í❦ì❭é➡í✰ë❛ô✍ò✟✩ õ✹ö♠ç✿è❦ò✥ö❫è❦í✆✘ í❦ò✥í❦ç❛õ ★ô❄ê ✰✡✮ ✪ ➎➇➆❫➏r➆❫Ð❭➉r➆❫➔ ➏❛➉r➆★Ó☎➊✵➏❛➋❋➏❛➂✚➈✵➒❸➋✥➪✖➈✍➄❛➏✂➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘❞➊➡➒❷➆❫➄❛➄✍➻ ➁➴➂✚➈✍➉❛➈☎➆✥➉❛➈✝✆ ➋♠➏❛➂✚➈✍➉❳➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘✪➒❸➈❄➚♠➈➡➒❸➄✂➆✥➶❩➋✥➚♠➈✂➏❛➂❁➈✹➎➇➆❫➏r➆✥Ð♠➉r➆❫➔ ➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘✪➒❸➈❄➚♠➈➡➒❦➻ ☎➴➏❢➏❛➂✚➈❳➄❛➈✍➉❛➚♠➃❸➊✍➈❳➃❸➌❤➏❛➈✍➉➍➑❏➆❫➊❄➈♠➹❤➪✖➈❳➅✚➉❛➋❫➚♠➃❸➎✚➈✞✆✾➪✖➃❸➎❁➈r➒❸➘✵➄❛➅❩➆✥➊✍➈✍➎✵➏r➆✥➉➍➓ Ð♠➈❄➏✯➎✚➈➡➒❷➆✍➘❞➶❩➋♠➐✚➌❁➎✚➄✝✟✖ ➑s➋♠➉✂➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎→➄❛➈✍➉❛➚♠➃❸➊✍➈✞➭❏➆❫➏☎➆✪➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉ ➄❛➪✖➃❸➏❛➊➡➂➇➺➡➻❋➁❳➂✚➈✆➅✚➉➍➃❸➋♠➉➍➃❸➏➍➃❸➈❄➄❋➆❫➉❛➈☎➐✚➄❛➈✍➎❞➏❛➋✪➄❛➈✍➅➇➆❫➉r➆✥➏❛➈✪➏❛➂✚➈☎➏❛➉r➆★Ó☎➊✵➑❦➋❭➉ ➏❛➂✚➈✯➎❤➃✇Þ☛➈✍➉❛➈✍➌❤➏✠✆❅➊➡➒❷➆✥➄❛➄❛➈✍➄✍➻❳➁❳➂❁➈❄➄❛➈✯➶❩➋❭➐❁➌✚➎✚➄✠✟✖ ➆✥➉❛➈☎↔❩➫❤↕❩➈✍➄❛➏➍➃❸➔☎➆✥➏❛➈✍➄ ➋❫➑❱➏❛➂❁➈➮➆❫➊✍➏❛➐➇➆★➒❱➎❁➈r➒✇➃❸➚♠➈✍➉❛➈✍➎❋➎✚➈➡➒❷➆✍➘❤➄✍➻✾✬❳➆✥➏❛➂❁➈✍➉✍➹➇➏❛➂❁➈❄➘☎➆❫➉❛➈✲➝➮Ï✰➜r➢③➫❤➜r➢❤➐❁➅❤➓ ➅❩➈✍➉❢➶➇➋♠➐✚➌❁➎✚➄✖➆✥➌❁➎✂➏❛➂✚➈✖➌❁➈❄➏➍➪✖➋❭➉❛✃✹➏❛➉➍➃❸➈✍➄✍➹♠➏❛➂✚➉❛➋♠➐✚Ð♠➂➾➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌☎➊✍➋❭➌❤➏❛➉❛➋❫➒ ➅❩➋✥➒✇➃❸➊➡➃❸➈✍➄✍➹✚➏❛➋✠✃❭➈✍➈✍➅✂➷❤➐❁➈✍➐✚➈➡➃❸➌✚Ð✹➎❁➈r➒❷➆★➘❤➄✦➆❫➏✰➈✥➆❫➊r➂✂➄❛➪Ô➃❸➏❛➊r➂✹➑❦➋♠➉❱➆Ô➅➇➆❫➉❛➏➍➃❸➊❄➐❼➓ ➒❷➆❫➉✿➊➡➒❷➆✥➄❛➄✟✗❩➪✖➈➡➒✇➒➇➶❩➈➡➒❸➋✥➪→➏❛➂❁➈✍➄❛➈❳➶❩➋❭➐❁➌✚➎✚➄✡✟✖ ➻✻➼✞➈❳➔✵➈✍➌❤➏➍➃❸➋❭➌❁➈✍➎☎➈✥➆✥➉➍➒✇➃❸➈✍➉ ➏❛➂➇➆❫➏Ò➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈✲➆❫➅✚➅❼➒✇➃❸➊★➆❫➏➍➃❸➋♠➌➇➄✹➂❩➆✍➚♠➈✠➒✇➃❸➔✂➃❸➏❛➄✠➏❛➋✵➏❛➂✚➈➡➃❸➉Ò➆❫➎➇➆❫➅✚➏r➆❫➶❤➃✇➒✇➃③➏❏➘ ✰ ➏❛➂✚➈✍➄❛➈✵➶❩➋❭➐❁➌✚➎❁➄☛✟✖ ➆❫➉❛➈✹➃❸➌✚➎❼➃❸➊✥➆✥➏➍➃❸➚♠➈❋➋✥➑✖➄❛➐✚➊➡➂❋➒✇➃❸➔✯➃❸➏❛➄✍➻✆☎✽➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎ ➄❛➈✍➉❛➚♠➃❸➊✍➈✵➱➇➋✥➪ ➃❸➄✲➆❫➄❛➄➍➃❸Ð❭➌❁➈❄➎✸➆☎➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘❋➒❸➈✍➚♠➈r➒Ò➆❫➏✹➈✥➆✥➊➡➂✞➄❛➪✖➃❸➏❛➊➡➂→➭❦➌❁➋❭➏ ➌✚➈✍➊✍➈✍➄❛➄r➆✥➉➍➃✇➒❸➘Ñ➏❛➂✚➈✯➄r➆✥➔✵➈✹➒❸➈✍➚♠➈➡➒❢➃❸➌❋➈❄➚♠➈✍➉❛➘✏➄❛➪Ô➃❸➏❛➊r➂➇➺❀✰☛➏❛➂❁➈✏➝PÏ✰➜r➢③➫❤➜r➢❁➎✚➈➡➓ ➒❷➆✍➘❋➶❩➋♠➐✚➌✚➎▼➆✥➎❁➚❭➈✍➉❛➏➍➃❸➄❛➈✍➎✞➏❛➋✪➆✂➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈✂➱❩➋✥➪⑤➃❸➄Ò➏❛➂❁➈✂➄❛➐❁➔ ➋❫➑✿➏❛➂✚➈✹➆✥➅❁➅✚➉❛➋♠➅✚➉➍➃❷➆❫➏❛➈☞✟✖ ➆✍➒❸➋♠➌✚Ð☎➏❛➂❁➈✹➅➇➆❫➏❛➂☛➻✿➁➴➂✚➈✹➎✚➈➡➒❷➆✍➘✪➶❩➋❭➐❁➌✚➎✞➆❫➎❤➓ ➚♠➈❄➉❛➏➍➃❸➄❛➈✍➎❋➏❛➋✵➆✹Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❋➄❛➈✍➉❛➚♠➃❸➊❄➈P➱➇➋✥➪✸➃❸➄➴➏❛➂✚➈✹➵❢➆✥➉❛➈✍✃❤➂❤➓✢✑✹➆✍➒✇➒❷➆❫Ð❭➈✍➉ ➶❩➋❭➐❁➌✚➎✧➻ ➁➴➂❤➃❸➄❢➄❛➊r➂❁➈✍➔✲➈✳➂➇➆❫➄❢➏❛➂✚➈✳➅✚➉❛➋♠➶❤➒❸➈✍➔✛➏❛➂➇➆❫➏✍➹❭➄➍➃❸➌✚➊✍➈✖➎✚➈➡➒❷➆✍➘P➃❸➄❱➆❫➎✚➎❼➃❸➏➍➃❸➚❭➈♠➹ ➆❫➄❛✃♠➃❸➌✚Ð✹➑❦➋♠➉✿➆✠➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✌✟✖ ➆✥➏✿➆✠Ð❫➃❸➚♠➈❄➌✵➄❛➪✖➃❸➏❛➊➡➂✵➎✚➋❤➈✍➄✖➌✚➋♠➏❢➎❤➃❸➉❛➈✍➊✍➏➍➒❸➘ ➔✵➈✥➆✥➌▼➏❛➂➇➆❫➏✍✟✖ ➃❸➄✵➏❛➂✚➈☎➏r➆❫➉❛Ð❭➈✍➏✯➎✚➈➡➒❷➆✍➘▼➶❩➋❭➐❁➌✚➎Ý➑s➋♠➉✹➏❛➂✚➈✆➅➇➆❫➏❛➂→➆✥➄✵➆ ➪❳➂❁➋✥➒❸➈♠➻❅✬❳➆❫➏❛➂✚➈✍➉✍➹✻➃❸➏✹➃❸➄✯➌✚➈✍➊✍➈❄➄❛➄r➆❫➉❛➘✞➏❛➋❋➆❫➎✚➎▼➐✚➅▼➏❛➂✚➈✵➏r➆❫➉❛Ð❭➈✍➏✯➎✚➈➡➒❷➆✍➘❤➄ ➆❫➏✲➈✥➆❫➊r➂❞➂❁➋❭➅✸➏❛➋❋➽❩➌✚➎✸➏❛➂✚➈✪➏r➆✥➉❛Ð♠➈✍➏✵➐✚➅✚➅❩➈✍➉☎➶❩➋♠➐✚➌✚➎✸➑❦➋♠➉✵➏❛➂✚➈✪➅➇➆❫➏❛➂☛➻ ➼✞➈✠➈✍➬❤➅❩➈❄➊✍➏✖➏❛➂✚➈➴➏❛➉❛➐✚➈➴Ï➇➫♠➥❄↕ ✟✍➝❤➤✥↕❦➫✳➶➇➋♠➐✚➌❁➎✚➄❳➋❫➚❭➈✍➉❳➆➴➒❸➋❭➌❁Ð✵➅➇➆❫➏❛➂☎➏❛➋✹➶❩➈ ➄➍➃❸Ð♠➌❤➃✇➽➇➊✥➆❫➌❤➏➍➒❷➘▼➒❸➋❫➪✦➈✍➉✯➏❛➂➇➆❫➌⑧➏❛➂❁➈✵➄❛➐✚➔❐➋❫➑❳➏❛➂✚➈✵➶❩➋❭➐❁➌✚➎✚➄✍✟✖ ➆❫➏✂➈✥➆❫➊r➂ ➂✚➋♠➅☛➻ ✎✳➐❁➏Ò➪✖➈✹➄❛➐✚Ð♠Ð♠➈❄➄❛➏✠➏❛➂➇➆❫➏✍➹❩➄➍➃❸➌✚➊✍➈✵➏❛➂❤➃❸➄➴➃❸➄➮➆❫➌✞➆✥➎❩➆✥➅❁➏➍➃❸➚❭➈✵➄❛➈✍➉❛➚♠➃❸➊✍➈♠➹ ➏❛➂✚➈✆➌✚➈✍➏❏➪✖➋♠➉❛✃✞➄❛➂❁➋❭➐❼➒❸➎→➌❁➋❭➏✵➆✥➏❛➏❛➈✍➔✵➅✚➏✵➏❛➋❋➊r➂❩➆✥➉r➆❫➊✍➏❛➈✍➉➍➃❸à❄➈✪➋♠➉✂➊✍➋❭➌❤➏❛➉❛➋❫➒ ➏❛➂✚➈✯➄❛➈✍➉❛➚♠➃❸➊❄➈✆➏❛➋☎Ð❭➉❛➈✥➆❫➏Ò➅❁➉❛➈✍➊r➃❸➄➍➃❸➋♠➌☛➹✖➆✥➌❁➎✞➏❛➂❤➐✚➄✠➄❛➂✚➋♠➐❤➒❸➎ ✲❛➐✚➄❛➏✠➐✚➄❛➈✵➏❛➂❁➈ ➄❛➐✚➔✾➋❫➑✿➏❛➂✚➈✠✟✖ ✳ ➄Ò➆❫➄❳➏❛➂✚➈✂➆❫➎✚➚♠➈✍➉❛➏➍➃❸➄❛➈❄➎Ñ➶❩➋❭➐❁➌✚➎☛➻ Ô➋❭➌❁➄➍➃❸➎✚➈✍➉✹➃❸➌❋➔✵➋♠➉❛➈✲➎❁➈✍➏r➆✍➃✇➒✿➂❁➋✥➪✽➏❛➂✚➈✯➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘Ñ➄❛➊r➂❁➈❄➔✵➈✯➪✦➋♠➉❛✃❤➄✍➻ ➯◗➑❱➏❛➂❁➈Ò➂❼➃❸Ð❭➂❁➈❄➄❛➏✠➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘✏➊➡➒❷➆❫➄❛➄❳➂➇➆❫➄Ò➆✹➔✵➋♠➔✵➈❄➌❤➏r➆❫➉❛➘☎➌❁➈❄➈✍➎✆➑❦➋❭➉✖➈✍➬❤➏❛➉r➆ ➶➇➆❫➌✚➎❁➪✖➃❸➎✚➏❛➂Ñ➎✚➐❁➈✂➏❛➋☎➆✯➶✚➐✚➉❛➄❛➏✠➶❤➘☎➄❛➈✍➚♠➈❄➉r➆★➒❢➋✥➑❢➃❸➏❛➄✠➄❛➋❭➐❁➉❛➊✍➈❄➄✍➹❩➃❸➏➴➄❛➏❛➈★➆★➒❸➄ ➏❛➂✚➈✵➶❩➆✥➌❁➎✚➪✖➃❸➎❁➏❛➂✞➑❦➉❛➋♠➔ ➏❛➂❁➈✂➒❸➋❫➪✦➈✍➉✹➊➡➒❷➆❫➄❛➄❛➈❄➄✍➻✵➁❳➂❁➈✂➌✚➈✍➬❤➏✹➊➡➒❷➆❫➄❛➄✂➏❛➂❤➐✚➄ ➄❛➈✍➈✍➄✏➆✥➄✪➆❋➶➇➆❫➄❛➈➡➒✇➃❸➌✚➈▼➋❫➑➮➋♠➅❩➈❄➉r➆❫➏➍➃❸➋♠➌✫➏❛➂✚➈❋➆❫Ð❭Ð♠➉❛➈✍Ðä➆❫➏❛➈ ✲➍➃❸➏❛➏❛➈✍➉☎➋❫➑➮➏❛➂❁➈ ➂❤➃❸Ð♠➂✚➈✍➉✠➊➡➒❷➆✥➄❛➄✍➻✿➁➴➂❤➃❸➄❳Ð♠➈✍➏❛➄✿➑❏➆❫➊✍➏❛➋❭➉❛➈✍➎✪➏❛➋❭Ð♠➈✍➏❛➂✚➈✍➉✖➪✖➃❸➏❛➂✏➏❛➂❁➈➮➆❫Ð♠Ð❭➉❛➈✍Ð❤➆✥➏❛➈ ➶✚➐❁➉❛➄❛➏➍➃❸➌✚➈✍➄❛➄❋➋❫➑➮➏❛➂❼➃❸➄✏➊➡➒❷➆❫➄❛➄✏➏❛➋Ñ➅✚➉❛➋❤➎✚➐❁➊✍➈✞➏❛➂✚➈❋➏❛➋♠➏r➆★➒✧✲➍➃❸➏❛➏❛➈✍➉✯➑s➋♠➉☎➏❛➂❁➈ ➄❛➈✍➊✍➋❭➌❁➎❞➊➡➒❷➆✥➄❛➄✍➻✞➁➴➂❤➃❸➄✂➊✥➆❫➄❛➊✥➆✥➎❁➈❄➄✵➎❁➋✥➪❳➌❞➏❛➋❋➏❛➂❁➈☎➎➇➆❫➏r➆✥Ð♠➉r➆❫➔❐➏❛➉r➆✍Ó☎➊♠➹ ➪❳➂❼➃❸➊r➂❋Ð♠➈❄➏❛➄✠➪❳➂❩➆✥➏❛➈✍➚♠➈❄➉✠➶➇➆❫➌✚➎❁➪✖➃❸➎✚➏❛➂ ➃❸➄✠➒❸➈➡➑s➏❛➋❫➚❭➈✍➉✹➆✥➌❁➎❋➄❛➐❤Þ☛➈✍➉❛➄❳➑❦➉❛➋♠➔ ➏❛➂✚➈✞➆❫➊✍➊✍➐✚➔✂➐❼➒❷➆✥➏❛➈✍➎ ✲❏➃❸➏❛➏❛➈✍➉✍➻✩☎✠➄☎➪✖➈❋➆✥➉❛Ð♠➐✚➈ ➒❷➆✥➏❛➈✍➉✍➹❳➏❛➂✚➈❋➎❩➆✥➏r➆❫Ð♠➉r➆✥➔ ➏❛➉r➆★Ó➾➊✯➄❛➂✚➋♠➐❤➒❸➎▼➅✚➉❛➋♠➶➇➆❫➶❤➒❸➘▼➶❩➈✂Ð❫➃❸➚♠➈❄➌❞➆❫➌✞➆✍➚♠➈❄➉r➆❫Ð♠➈✲➉r➆❫➏❛➈✂➋❫➑Ô➆❫➏➴➒❸➈★➆❫➄❛➏ ✂✔✕✄✁⑤➋♠➉✿➄❛➋❁➹❭➶❩➋❭➏❛➂✵➏❛➋➴➃❸➌✚➄❛➐❁➉❛➈Ò➏❛➂❩➆✥➏✻➃❸➏✳➔➾➆❫✃♠➈❄➄✳➄❛➋❭➔✵➈✖➅✚➉❛➋♠Ð♠➉❛➈❄➄❛➄✳➋❭➌✵➏❛➂❁➈ ➆✍➚♠➈✍➉r➆✥Ð♠➈☎➆✥➌❁➎✞➏❛➋✪➅✚➉❛➋✥➚♠➃❸➎❁➈✏➆✵➉❛➈✥➆✥➄❛➋♠➌➇➆❫➶❤➒❸➈✪➅❩➋❤➋✥➒✖➋✥➑✖➶➇➆❫➌✚➎❁➪✖➃❸➎✚➏❛➂Ñ➑❦➋❭➉ ➏❛➂✚➈➴➂❤➃❸Ð♠➂✚➈✍➉✖➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘☎➏❛➉r➆★Ó☎➊❳➏❛➋✹➶❩➋❭➉❛➉❛➋❫➪❞➑❦➉❛➋♠➔✽➎✚➐❁➉➍➃❸➌✚Ð✂➔✵➋♠➔✵➈❄➌❤➏r➆❫➉❛➘ ➋✥➚♠➈✍➉➍➒❸➋ä➆❫➎✚➄✍➻ ✛❁➋♠➉❳➆➴➒❸➋✥➪✖➈❄➉✖➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✏➊➡➒❷➆❫➄❛➄✍➹➇➪➴➂➇➆❫➏✿➄❛➋♠➐✚➉❛➊✍➈✹➋✥➑ ✲❏➃❸➏❛➏❛➈✍➉✖➪✖➃✇➒✇➒✰➎❁➋❭➔✯➓ ➃❸➌➇➆❫➏❛➈✹➃❸➏❛➄➴➋❭➶❁➄❛➈❄➉❛➚♠➈✍➎❋➶❩➈❄➂❩➆✍➚♠➃❸➋♠➉✍❒✿➃❸➏❛➄✖➃❸➌❤➏❛➉➍➃❸➌❁➄➍➃❸➊✏➆❫Ð❭Ð♠➉❛➈✍Ðä➆❫➏❛➈✹➶❩➈❄➂❩➆✍➚♠➃❸➋♠➉ ➋♠➉Ò➏❛➂❁➈✾✲➍➃❸➏❛➏❛➈✍➉Ò➄❛➂❼➃✇➑s➏❛➈✍➎ ➑s➉❛➋♠➔ ➏❛➂❁➈✂➂❼➃❸Ð❭➂❁➈❄➉✹➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✞➊➡➒❷➆❫➄❛➄❛➈✍➄✄✸➯❦➑✿➏❛➂❁➈ ➏r➆❫➉❛Ð❭➈✍➏✵Ð❭➋❤➆★➒❸➄✵➑❦➋❭➉❂✲❏➃❸➏❛➏❛➈✍➉☎➆❫➉❛➈✏➪✖➃❸➎❁➈➡➒❸➘❞➄❛➅➇➆❫➊✍➈❄➎⑤➭❏➆❫➌✚➎▼➑❦➋♠➉✲➏❛➂❁➈✏➅❁➐✚➉➍➓ ➅❩➋❭➄❛➈✠➋❫➑❱➉❛➋♠➐✚Ð♠➂☎➈✍➄❛➏➍➃❸➔☎➆✥➏➍➃❸➋♠➌✞➆✥➄✖➪✖➈❳➄❛➐❁Ð❭Ð♠➈✍➄❛➏❛➈❄➎❋➆❫➶➇➋❫➚❭➈✠➏❛➂✚➈✍➘✵➅✚➉❛➋♠➶➇➆★➓ ➶❤➒❸➘✪➌✚➈✍➈✍➎✪➶➇➈✹➌❁➋✵➊r➒❸➋♠➄❛➈✍➉Ò➏❛➂❩➆✥➌❋➆❫➌✏➋♠➉❛➎✚➈✍➉❳➋❫➑❱➔☎➆❫Ð❭➌❼➃❸➏❛➐✚➎❁➈★➺✠➏❛➂❁➈✍➌✏➏❛➂❁➈ ➈✍➬❤➅❩➋❭➉❛➏❛➈✍➎❋✲➍➃❸➏❛➏❛➈✍➉✠➑❦➉❛➋❭➔❅➏❛➂✚➈✆➂❤➃❸Ð♠➂✚➈✍➉✂➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘❞➊➡➒❷➆✥➄❛➄✯➄❛➂✚➋♠➐❤➒❸➎❞➶❩➈✏➆❫➌ ➋♠➉❛➎✚➈✍➉✵➋✥➑✹➔☎➆✥Ð♠➌❤➃❸➏❛➐❁➎✚➈✪➒❸➈✍➄❛➄☎➏❛➂➇➆❫➌❞➏❛➂❁➈☎➃❸➌❤➏❛➉➍➃❸➌❁➄➍➃❸➊✞➶❩➈✍➂➇➆✍➚♠➃❸➋♠➉✏➋❫➑Ò➏❛➂❁➈ ➊➡➒❷➆✥➄❛➄✍➹Ò➆❫➌✚➎✸➏❛➂✚➈❋➊➡➒❷➆❫➄❛➄❛➈✍➄☎➄❛➂✚➋♠➐❤➒❸➎⑤➐✚➄❛➐➇➆★➒✇➒❸➘✸➋❭➅❩➈✍➉r➆✥➏❛➈❋➔✵➋♠➉❛➈❋➋❭➉✯➒❸➈✍➄❛➄ ➃❸➌✚➎❁➈✍➅➇➈✍➌❁➎✚➈✍➌❤➏➍➒❷➘♠➻✹➁❳➂❤➐✚➄✍➹☛➆✵➅❩➆✥➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✂➊➡➒❷➆❫➄❛➄❳➃❸➄➴➃❸➄❛➋❫➒❷➆✥➏❛➈✍➎Ñ➑❦➉❛➋❭➔✾➏❛➂❁➈ ➒❸➋✥➪✖➈✍➉✵➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘✸➊➡➒❷➆✥➄❛➄❛➈✍➄☎➶❤➘▼➏❛➂✚➈❋➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘❞➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð❑➆✍➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔ ➆❫➌✚➎❋➃❸➄➴➃❸➌✞➈➡Þ☛➈✍➊✍➏❳➃❸➄❛➋❫➒❷➆✥➏❛➈✍➎Ñ➑❦➉❛➋❭➔✾➏❛➂❁➈✂➂❼➃❸Ð❭➂❁➈❄➉✹➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✞➊➡➒❷➆❫➄❛➄❛➈✍➄✹➶➇➈➡➓ ➊✥➆✥➐❁➄❛➈☎➏❛➂✚➈➡➃❸➉❃✲➍➃❸➏❛➏❛➈✍➉✹➪✖➃✇➒✇➒❳➶➇➈✆➄❛➋❋➔✂➐✚➊➡➂▼➄❛➔➾➆★➒✇➒❸➈✍➉✵➏❛➂➇➆❫➌⑧➏❛➂❩➆✥➏✯➋✥➑❳➏❛➂❁➈ ➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✹➊➡➒❷➆✥➄❛➄✍➻ ➼Ñ➈✹➂➇➆✍➚♠➈✂➄➍➃❸➔✯➐❤➒❷➆❫➏❛➈✍➎✞➏❛➂❤➃❸➄➴➐✚➌❤➃✇➽❩➈✍➎✞➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✞➆✍➒❸Ð♠➋♠➉➍➃❸➏❛➂✚➔❅➐✚➄➍➓ ➃❸➌✚Ð❋➏❛➂❁➈☎➄r➆❫➔✲➈✆➄➍➃❸➔✂➐❤➒❷➆❫➏➍➃❸➋♠➌→➊✍➋♠➌❤➽➇Ð♠➐✚➉r➆❫➏➍➃❸➋♠➌ ➆❫➄✹➐✚➄❛➈✍➎Ñ➑❦➋❭➉✹➁✿➆❫➶❤➒❸➈❅✮❼➹ ➏❛➂➇➆❫➏✹➂➇➆❫➄ ✮✒✮✆➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✵➱❩➋✥➪❳➄✹➪✖➃❸➏❛➂Ñ➃❸➎✚➈✍➌❤➏➍➃❸➊✥➆★➒Ô➄❛➏r➆❫➏➍➃❸➄❛➏➍➃❸➊✥➆✍➒✠Ð❭➈✍➌✚➈✍➉➍➓ ➆❫➏➍➃❸➋❭➌Ñ➅✚➉❛➋❤➊✍➈✍➄❛➄❛➈❄➄✹➶❁➐✚➏✹➪❳➂❼➃❸➊r➂Ñ➏❛➉r➆✍➚❭➈✍➉❛➄❛➈✯➎❤➃✇Þ☛➈✍➉❛➈❄➌❤➏✹➌❁➈✍➏❏➪✖➋♠➉❛✃❋➅➇➆❫➏❛➂✚➄✍➻ ➁❢➋✂➏❛➂❁➈✍➄❛➈❉✮❏✮P➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✹➱❩➋✥➪❳➄✖➪✖➈➮➆❫➎✚➎❁➈✍➎ ✮✹➎➇➆❫➏r➆❫Ð❭➉r➆❫➔✾➁❳➵▼➊✍➋♠➌❤➓ ➌✚➈✍➊✍➏➍➃❸➋♠➌✚➄✍➻ ➯➍➌✸➏❛➂❤➃❸➄✪➄➍➃❸➔✂➐❤➒❷➆❫➏➍➃❸➋♠➌✰➹✠✗✞➋❫➑✂➏❛➂✚➈Ñ➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈ ➱➇➋✥➪❳➄❋➆❫➉❛➈ Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎✏➄❛➈✍➉❛➚♠➃❸➊✍➈✹➊➡➒✇➃❸➈✍➌❤➏❛➄☛✰ ✝✠➋❫➑✰➏❛➂✚➈✍➄❛➈✠➂➇➆✍➚♠➈✹➆✹➊➡➒❸➋❤➊r✃✵➉r➆❫➏❛➈Ò➈✍➷❤➐➇➆★➒
path delay measure PG‖ type path delay measur length n g verage 5f.3522./922.6676Eow 0厂0.QU88 35.020Q255.f658.2Low Table e mean, 99.9'th percentile time)in eight sample fo nified scheda g delays(measured in milliseconds, which is a single packet m. The network configuration is shown in Figure Q Each inter-switch link is utilized over to their peak packet generation rate (we denote such flows loads, and much of the novelty of our unified scheduling by Guaranteed-Peak) and the ot her 2 have a clock rate equal algorithm is our provision for predicted service, which can to their average packet generation rat denote such floy by Guaranteed-g verage). See the g ppendix for details on he statistical generation process and the values of the ay We have now completed the first parts of our archite erage and peak rates. The remaining g real-time flows are ture. We have described a model for the low-level packet redicted service clients served by two priority classes, 7 forwarding algorithm, which is a sharing discipline inside an lows are in the high priority class(we denote such flows by isolation discipline redicted-High) and the other 0 fows are in the low pri ample of such a scheme, which provides both of our service rity class ch flows by Predicted-Low). If we commit ment models, guaranteed and predicted. The scheme ook at the traffic traversing each link, it consists of one data- rovides several predicted service classes with different delay ram connection and (O real-time flows: 2 Guaranteed-Pea bounds, and uses a particular technique(FIFO+) to provide laranteed-g verage, 3 Predicted-High, and f Pi low jitter, and to provide a jitter bound that does not vary strongly with the number of hops in the path Results for eight sample real-time flows in this simulation are presented in Table 3. The mean, 99.9'th percentile, and he maximum delays are displayed for all realtime flows and the Parekh-Gallager (P-g)delay bounds are shown gs a part of the definition of the unified scheduling algo- he guaranteed service flows. We see that all of the ithm, we have also defined our service interface. In fact teed service flows received worst-case delays th for the e interface, one for gua within the Parekh-Gallager bounds. Not surprisingly, the anteed service and another for predicted service Guaranteed-Peak flows experienced much lower delays than For guaranteed service, the interface is simple: the source only needs to specify the needed High flows experienced lower delays than the Predicted-Low twork t his rate. The uses It For the load pattern described here, the de ys of the Predicted-High fows were lower than those of () If the delay is unsuitable, it must higher clock he comparable Guaranteed-Peak flows, and the delays of etwork does no con formance chee he predicted-Low flows were lower than those of the com guaranteed service flow, because the flow does not make any parable Guaranteed-g verage fows; howe ver, this relation traffic characterization commitment to the net work bet ween the delays of the two classes is an artifact of the For predicted service, the service interface must charac art icular load pat tern and is not necessarily indicative of a terize both the traffic and th ice. For the characteriza general patte Table 3 is the performance of the d (r, b) of the token bucket traffic filter to which it claims its gram traffic which experienced a very low drop rate, around traffic will conform. Note that in the guaranteed case the 0. Co. The overall utilization of the net work was over 99%o, client did not need to inform the network of its bucket size& ith 83.5% of this being real-time traffic. It is im portant te b. Separately, the source must request the needed service note that if all of th This involves selecting a suit able delay D and a target loss eed service with a clock rate rate z the application can tolerate. The network will use date many fe these numbers to assign the source to an aggregate class at the utilization due to realtime traffic would be reduced te viding predicted service allows the vice, the parameters of the service interface are the filter rate k to operate with a higher degree of real-time traffic and size(r, b)and the delay and loss characteristics(D, z) rovide predicted force the traffic commit ments made by the clients. Enforce incomplete, are qualit atively consistent with our analysis. I e ently attempting to more fully valid ate is checked at the edge of the network (i.e, the first switch through simula progress in a subsequent publication. Note that much of the This assumes that the switches al ways hawe sufficient buffering hallenge here is de termining how to evaluate out l'oa1-time the buffers in cases of overflow knowled equired for the switch to manage There is
✑✠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎✞➲❼➈❄➉❛➚♠➃❸➊✍➈ ➵❢➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎✞➲❼➈❄➉❛➚♠➃❸➊✍➈ ➏➍➘❤➅➇➈ ➅❩➆✥➏❛➂ ➎✚➈➡➒❷➆✍➘✪➔✲➈✥➆❫➄❛➐✚➉❛➈ ➵✻➓✢✑ ➏➍➘❤➅❩➈ ➅➇➆❫➏❛➂ ➎✚➈➡➒❷➆✍➘❋➔✵➈✥➆✥➄❛➐❁➉❛➈ ➒❸➈❄➌❁Ð❭➏❛➂ ➔✲➈✥➆❫➌ ✓✖✓❤➻✼✓✂✁✹➃✇➒❸➈ ➔☎➆✥➬ ➶❩➋♠➐✚➌❁➎ ➒❸➈✍➌✚Ð♠➏❛➂ ➔✲➈✥➆❫➌ ✓✬✓❼➻ ✓✂✁P➃✇➒❸➈ ➔☎➆✥➬ ➵✻➈✥➆❫✃ ✟ ✣ ➻✼✓❏✮ ✂✏✟✚➻✼✝●✂ ✂✭✗❼➻ ✓✲✟ ✮✖✝❼➻ ✗✖✝ ßÔ➃❸Ð❭➂ ✟ ✝❼➻✒✂✟✮ ✧❼➻ ✮✾✂ ✂✭✕❼➻ ✣ ✂ ➵✻➈✥➆❫✃ ✮ ✝❤➻✼✧✬✕ ✧❼➻✒✂✟✮ ✧❤➻ ✣ ✓ ✂✖✂♠➻ ✣ ✙ ßÔ➃❸Ð❭➂ ✮ ✂♠➻ ✙✾✂ ✗❼➻ ✧✖✓ ✣ ➻✼✕✔✟ ☎Ò➚♠➈✍➉r➆✥Ð♠➈ ✝ ✗✲✟✚➻✼✝✬✗ ✮✲✟✻✮❤➻ ✟❨✓ ✮✬✓✒✮❤➻ ✟❨✙ ✙✾✂✖✂❭➻ ✣ ✙ ✳✧➋✥➪ ✝ ✂✔✧❤➻✼✕✔✟ ✂✔✕✖✕❤➻✼✗✬✗ ✂✟✮✬✕❼➻ ✧✖✧ ☎Ò➚♠➈✍➉r➆✥Ð♠➈ ✂ ✝✖✗❤➻ ✟❨✕ ✮✾✂ ✣ ➻✼✓●✂ ✮✬✗✖✗❤➻ ✟❨✙ ✗✖✧✖✧❤➻ ✮✔✟ ✳✧➋✥➪ ✂ ✙❼➻ ✓✲✟ ✣ ✮❼➻ ✟●✙ ✂✔✕✬✙❼➻ ✗✖✙ ➁✿➆✥➶❼➒❸➈ ✝❼❒✂➁❳➂❁➈☎➔✵➈★➆❫➌☛➹❏✓✬✓❼➻ ✓ ✳ ➏❛➂Ñ➅➇➈✍➉❛➊✍➈✍➌❤➏➍➃✇➒❸➈♠➹Ò➆✥➌❁➎⑧➔☎➆❫➬♠➃❸➔✂➐❁➔ ➷❤➐❁➈❄➐❁➈➡➃❸➌✚Ð▼➎✚➈➡➒❷➆✍➘❤➄✪➭❦➔✵➈✥➆✥➄❛➐❁➉❛➈❄➎①➃❸➌❞➔✯➃✇➒✇➒✇➃❸➄❛➈❄➊✍➋♠➌✚➎❩➄✍➹❳➪❳➂❼➃❸➊r➂▼➃❸➄✵➆✪➄➍➃❸➌✚Ð❫➒❸➈❋➅➇➆❫➊r✃♠➈❄➏ ➏❛➉r➆✥➌❁➄❛➔✂➃❸➄❛➄➍➃❸➋♠➌✪➏➍➃❸➔✲➈✥➺❱➃❸➌☎➈➡➃❸Ð❭➂❤➏✳➄r➆✥➔✵➅❤➒❸➈➴➱➇➋❫➪❳➄✿➐❁➌✚➎✚➈✍➉✳➏❛➂✚➈❳➐❁➌❤➃✇➽➇➈✍➎☎➄❛➊➡➂✚➈✍➎✚➐❼➒✇➃❸➌✚Ð✏➆★➒❸Ð♠➋❭➉➍➃❸➏❛➂❁➔✏➻✖➁❳➂❁➈✦➌❁➈✍➏❏➪✖➋♠➉❛✃✂➊✍➋❭➌❼➽➇Ð♠➐✚➉r➆❫➏➍➃❸➋❭➌✆➃❸➄✖➄❛➂✚➋✥➪➴➌✂➃❸➌✆✛✧➃❸Ð♠➐✚➉❛➈ ✂❭➻ ✎✳➆✥➊➡➂✪➃❸➌❤➏❛➈✍➉➍➓❏➄❛➪✖➃❸➏❛➊r➂✪➒✇➃❸➌❁✃✪➃❸➄❳➐✚➏➍➃✇➒✇➃❸à✍➈✍➎▼➋✥➚♠➈❄➉ ✓✬✓✁✆➻ ➏❛➋✏➏❛➂❁➈r➃❸➉✹➅❩➈✥➆❫✃❋➅➇➆❫➊r✃♠➈❄➏✯Ð❭➈✍➌❁➈❄➉r➆❫➏➍➃❸➋♠➌⑧➉r➆❫➏❛➈✪➭❦➪✖➈✂➎✚➈✍➌❁➋❭➏❛➈✵➄❛➐✚➊➡➂❋➱❩➋✥➪❳➄ ➶❤➘✵✑✠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❤➓➳➵✻➈✥➆❫✃✚➺✖➆✥➌❁➎✂➏❛➂✚➈✖➋♠➏❛➂✚➈✍➉✌✮✖➂❩➆✍➚❭➈✠➆✖➊➡➒❸➋❤➊r✃✂➉r➆❫➏❛➈✖➈✍➷❤➐➇➆★➒ ➏❛➋Ò➏❛➂❁➈r➃❸➉✿➆★➚❭➈✍➉r➆❫Ð❭➈✖➅➇➆❫➊r✃♠➈✍➏❢Ð❭➈✍➌❁➈❄➉r➆❫➏➍➃❸➋♠➌✲➉r➆❫➏❛➈✠➭❦➪✖➈✿➎✚➈✍➌❁➋❭➏❛➈✖➄❛➐✚➊➡➂✹➱❩➋✥➪❳➄ ➶❤➘✤✑✠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎❼➓☎Ò➚♠➈✍➉r➆✥Ð♠➈✥➺➡➻⑧➲❤➈✍➈☎➏❛➂❁➈✆☎✠➅✚➅❩➈✍➌✚➎❤➃❸➬①➑s➋♠➉✹➎✚➈✍➏r➆★➃✇➒❸➄☎➋♠➌ ➏❛➂✚➈✵➄❛➏r➆✥➏➍➃❸➄❛➏➍➃❸➊✥➆★➒❳Ð❭➈✍➌✚➈✍➉r➆❫➏➍➃❸➋❭➌Ý➅✚➉❛➋❤➊✍➈✍➄❛➄✲➆❫➌✚➎✞➏❛➂❁➈✵➚❭➆✍➒❸➐❁➈✍➄✲➋❫➑✦➏❛➂❁➈☎➆✍➚♠➓ ➈❄➉r➆❫Ð♠➈✲➆❫➌✚➎❋➅❩➈✥➆✥✃✪➉r➆✥➏❛➈✍➄✍➻✿➁➴➂✚➈✹➉❛➈✍➔☎➆✍➃❸➌❼➃❸➌✚Ð✪✂ ✣ ➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✯➱➇➋✥➪❳➄✠➆✥➉❛➈ ➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✛➄❛➈❄➉❛➚♠➃❸➊✍➈▼➊➡➒✇➃❸➈✍➌❤➏❛➄✞➄❛➈✍➉❛➚❭➈✍➎❑➶❤➘✸➏❏➪✖➋❞➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✫➊➡➒❷➆❫➄❛➄❛➈✍➄✍➹ ✣ ➱➇➋✥➪❳➄✠➆✥➉❛➈➴➃❸➌❋➏❛➂✚➈✠➂❤➃❸Ð♠➂❋➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘✪➊➡➒❷➆✥➄❛➄✯➭❦➪✖➈Ò➎❁➈✍➌✚➋♠➏❛➈✹➄❛➐✚➊r➂✆➱➇➋❫➪❳➄❳➶❤➘ ➵❢➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓➳ßÔ➃❸Ð♠➂➇➺☎➆❫➌✚➎▼➏❛➂✚➈✵➋♠➏❛➂✚➈✍➉✆✂✔✕➾➱➇➋✥➪❳➄✵➆✥➉❛➈✯➃❸➌▼➏❛➂✚➈➾➒❸➋✥➪ ➅❁➉➍➃✇➓ ➋❭➉➍➃❸➏➍➘❋➊➡➒❷➆✥➄❛➄✆➭❦➪✦➈✯➎✚➈✍➌✚➋♠➏❛➈✵➄❛➐✚➊➡➂❋➱➇➋❫➪❳➄✹➶❤➘✞➵❢➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓✢✳☛➋✥➪✠➺➡➻✂➯❦➑✖➪✖➈ ➒❸➋❤➋❭✃✂➆❫➏✰➏❛➂❁➈✿➏❛➉r➆★Ó☎➊✿➏❛➉r➆✍➚♠➈✍➉❛➄➍➃❸➌✚Ð✹➈✥➆✥➊➡➂✠➒✇➃❸➌✚✃❩➹♠➃❸➏❢➊✍➋❭➌❁➄➍➃❸➄❛➏❛➄❢➋✥➑❩➋❭➌❁➈✿➎❩➆✥➏r➆★➓ Ð❭➉r➆❫➔✛➊❄➋♠➌✚➌❁➈✍➊❄➏➍➃❸➋♠➌☎➆✥➌❁➎✆✂✔✕✖➉❛➈✥➆✍➒✇➓➍➏➍➃❸➔✲➈✳➱❩➋✥➪❳➄✍❒✌✮✁✑✠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎❼➓➳➵✻➈✥➆✥✃❩➹ ✂✜✑✠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❤➓☎Ò➚♠➈✍➉r➆✥Ð♠➈♠➹ ✝✆➵❢➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓➳ßÔ➃❸Ð♠➂☛➹✦➆❫➌✚➎ ✟✏➵❢➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓ ✳☛➋✥➪P➻ ✬✖➈✍➄❛➐❤➒❸➏❛➄ ➑❦➋❭➉✻➈➡➃❸Ð❭➂❤➏✳➄r➆✥➔✵➅❤➒❸➈✖➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✖➱➇➋✥➪➴➄✻➃❸➌✂➏❛➂❤➃❸➄❢➄➍➃❸➔✯➐❤➒❷➆❫➏➍➃❸➋❭➌ ➆✥➉❛➈✹➅❁➉❛➈❄➄❛➈✍➌❤➏❛➈✍➎✆➃❸➌✏➁✿➆❫➶❤➒❸➈✵✝❼➻✿➁❳➂❁➈Ò➔✵➈✥➆✥➌✧➹ ✓✖✓❼➻ ✓ ✳ ➏❛➂☎➅❩➈✍➉❛➊✍➈❄➌❤➏➍➃✇➒❸➈♠➹❢➆✥➌❁➎ ➏❛➂✚➈❋➔☎➆✥➬♠➃❸➔✯➐✚➔ ➎✚➈➡➒❷➆✍➘❤➄❋➆✥➉❛➈✪➎❤➃❸➄❛➅❼➒❷➆✍➘❭➈✍➎ã➑s➋♠➉✏➆★➒✇➒➮➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✪➱➇➋✥➪➴➄✍➹ ➆✥➌❁➎❋➏❛➂✚➈✂➵❢➆❫➉❛➈❄✃❤➂❼➓■✑✹➆★➒✇➒❷➆✥Ð♠➈✍➉✵➭❏➵✻➓■✑✹➺✖➎❁➈➡➒❷➆✍➘✏➶❩➋♠➐✚➌❁➎✚➄✹➆✥➉❛➈✹➄❛➂❁➋✥➪❳➌✆➑s➋♠➉ ➏❛➂✚➈✹Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❋➄❛➈✍➉❛➚♠➃❸➊❄➈✠➱❩➋✥➪❳➄✍➻✿➼Ñ➈✹➄❛➈✍➈Ò➏❛➂❩➆✥➏Ò➆★➒✇➒❢➋✥➑❢➏❛➂✚➈✠Ð❭➐❩➆✥➉r➆❫➌❤➓ ➏❛➈❄➈✍➎❋➄❛➈✍➉❛➚♠➃❸➊✍➈✵➱➇➋❫➪❳➄Ò➉❛➈✍➊✍➈➡➃❸➚❭➈✍➎Ñ➪✦➋♠➉❛➄❛➏➍➓❏➊✥➆❫➄❛➈✂➎✚➈➡➒❷➆✍➘❤➄✹➏❛➂❩➆✥➏✹➪✖➈✍➉❛➈✹➪✖➈➡➒✇➒ ➪✖➃❸➏❛➂❤➃❸➌❑➏❛➂❁➈ ➵❢➆❫➉❛➈✍✃❤➂❤➓■✑✹➆★➒✇➒❷➆✥Ð♠➈✍➉✏➶❩➋♠➐✚➌❁➎✚➄✍➻ ➸➴➋♠➏✲➄❛➐❁➉❛➅✚➉➍➃❸➄➍➃❸➌❁Ð✥➒③➘❭➹✠➏❛➂✚➈ ✑✠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❤➓➳➵✻➈✥➆❫✃☎➱❩➋✥➪❳➄✖➈✍➬❤➅❩➈❄➉➍➃❸➈✍➌❁➊❄➈✍➎❋➔✯➐✚➊r➂✆➒❸➋✥➪✖➈✍➉✖➎✚➈➡➒❷➆✍➘❤➄✠➏❛➂➇➆❫➌ ➏❛➂✚➈✚✑✠➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎❼➓☎✠➚❭➈✍➉r➆❫Ð❭➈Ý➱➇➋✥➪➴➄✍➻ ➲♠➃❸➔✯➃✇➒❷➆✥➉➍➒❸➘♠➹✪➏❛➂✚➈❞➵❢➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓ ß✖➃❸Ð♠➂✵➱➇➋❫➪❳➄✿➈✍➬❤➅❩➈✍➉➍➃❸➈❄➌❁➊✍➈❄➎✆➒❸➋✥➪✖➈✍➉✿➎❁➈r➒❷➆★➘❤➄✦➏❛➂❩➆✥➌✵➏❛➂❁➈Ò➵❢➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓✢✳☛➋✥➪ ➱➇➋✥➪❳➄✍➻✩✛❁➋♠➉✲➏❛➂❁➈✏Ð❫➃❸➚♠➈❄➌✸➒❸➋❤➆❫➎❑➅➇➆❫➏❛➏❛➈❄➉❛➌✸➎✚➈✍➄❛➊✍➉➍➃❸➶❩➈✍➎❑➂✚➈✍➉❛➈♠➹❳➏❛➂✚➈✪➎✚➈➡➓ ➒❷➆✍➘❤➄☎➋❫➑➮➏❛➂❁➈✞➵❢➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓➳ßÔ➃❸Ð♠➂✸➱➇➋✥➪❳➄✆➪✦➈✍➉❛➈✵➒❸➋❫➪✦➈✍➉☎➏❛➂❩➆✥➌✸➏❛➂✚➋♠➄❛➈✪➋✥➑ ➏❛➂✚➈✪➊❄➋♠➔✵➅➇➆❫➉r➆✥➶❼➒❸➈✤✑✠➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎❼➓➳➵✻➈✥➆✥✃Ý➱➇➋❫➪❳➄✍➹Ô➆❫➌✚➎✸➏❛➂✚➈✪➎✚➈➡➒❷➆✍➘❤➄✆➋✥➑ ➏❛➂✚➈✲➵ ➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❤➓■✳✧➋✥➪✛➱❩➋✥➪❳➄✠➪✦➈✍➉❛➈P➒❸➋✥➪✖➈❄➉✠➏❛➂➇➆❫➌✞➏❛➂❁➋❭➄❛➈✵➋❫➑✿➏❛➂✚➈✯➊❄➋♠➔✯➓ ➅➇➆❫➉r➆✥➶❼➒❸➈✚✑✠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎❤➓☎Ò➚♠➈✍➉r➆✥Ð♠➈▼➱➇➋❫➪❳➄☛✰✂➂✚➋❫➪✦➈✍➚♠➈✍➉✍➹➮➏❛➂❼➃❸➄❋➉❛➈➡➒❷➆✥➏➍➃❸➋♠➌ ➶➇➈✍➏➍➪✖➈❄➈✍➌❞➏❛➂❁➈☎➎✚➈➡➒❷➆✍➘❤➄✆➋✥➑Ò➏❛➂❁➈✏➏➍➪✖➋❋➊➡➒❷➆✥➄❛➄❛➈✍➄✯➃❸➄☎➆✥➌→➆❫➉❛➏➍➃✇➑❏➆✥➊✍➏✵➋❫➑Ò➏❛➂✚➈ ➅➇➆❫➉❛➏➍➃❸➊❄➐❼➒❷➆✥➉➴➒❸➋❤➆✥➎✪➅➇➆❫➏❛➏❛➈✍➉❛➌✏➆✥➌❁➎☎➃❸➄✖➌❁➋❭➏✖➌✚➈✍➊✍➈✍➄❛➄r➆✥➉➍➃✇➒❸➘✆➃❸➌✚➎❼➃❸➊✥➆✥➏➍➃❸➚♠➈✵➋✥➑❱➆ Ð❭➈✍➌✚➈✍➉r➆★➒✿➅➇➆❫➏❛➏❛➈✍➉❛➌☛➻ ➸➴➋❭➏✵➄❛➂✚➋❫➪❳➌❞➃❸➌✸➁✿➆❫➶❤➒❸➈✞✝ ➃❸➄☎➏❛➂✚➈✪➅❩➈✍➉➍➑s➋♠➉❛➔☎➆✥➌❁➊✍➈❋➋✥➑✹➏❛➂✚➈❋➎❩➆✥➏r➆★➓ Ð❭➉r➆❫➔ ➏❛➉r➆★Ó☎➊❳➪❳➂❼➃❸➊r➂☎➈✍➬❤➅❩➈✍➉➍➃❸➈❄➌❁➊✍➈❄➎✞➆✠➚♠➈✍➉❛➘✯➒❸➋✥➪❑➎✚➉❛➋♠➅☎➉r➆❫➏❛➈❭➹✚➆❫➉❛➋❭➐❁➌✚➎ ✕❤➻✰✂ ✁✆➻❳➁❳➂❁➈✂➋✥➚♠➈✍➉r➆✍➒✇➒✖➐✚➏➍➃✇➒✇➃❸à✥➆✥➏➍➃❸➋♠➌❞➋❫➑✦➏❛➂❁➈✂➌✚➈✍➏➍➪✖➋❭➉❛✃✪➪❳➆✥➄❳➋❫➚❭➈✍➉✌✓✬✓ ✁✆➹ ➪✖➃❸➏❛➂ ✧✖✝❼➻ ✗ ✁✛➋✥➑❢➏❛➂❤➃❸➄➴➶➇➈➡➃❸➌❁Ð✲➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✹➏❛➉r➆✍Ó☎➊♠➻✿➯➍➏❢➃❸➄✖➃❸➔✵➅❩➋♠➉❛➏r➆✥➌❤➏❳➏❛➋ ➌✚➋♠➏❛➈✹➏❛➂➇➆❫➏✖➃✇➑✦➆✍➒✇➒❢➋❫➑✿➏❛➂✚➈✠➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✹➱❩➋✥➪❳➄❳➂❩➆✥➎❋➉❛➈✍➷❤➐❁➈❄➄❛➏❛➈✍➎✪Ð❭➐❩➆✥➉r➆❫➌❤➓ ➏❛➈❄➈✍➎Ñ➄❛➈❄➉❛➚♠➃❸➊✍➈☎➪Ô➃❸➏❛➂⑧➆✵➊➡➒❸➋❤➊r✃Ñ➉r➆✥➏❛➈✯➈✍➷❤➐➇➆★➒✦➏❛➋✪➏❛➂✚➈➡➃❸➉✹➅❩➈★➆❫✃❋➉r➆❫➏❛➈❭➹✻➏❛➂✚➈ ➌✚➈✍➏❏➪✖➋♠➉❛✃✂➊✍➋♠➐❤➒❸➎☎➆✥➊✍➊✍➋♠➔✲➔✵➋❤➎❩➆✥➏❛➈✠➔☎➆✥➌❤➘P➑s➈✍➪✖➈✍➉❢➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈✖➱❩➋✥➪❳➄✖➆✥➌❁➎ ➏❛➂✚➈✵➐✚➏➍➃✇➒✇➃❸à✥➆❫➏➍➃❸➋❭➌✸➎✚➐❁➈✂➏❛➋✪➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✵➏❛➉r➆✍Ó☎➊✵➪✖➋❭➐❼➒❸➎✞➶❩➈✵➉❛➈✍➎✚➐✚➊✍➈✍➎✞➏❛➋ ➉❛➋❭➐❁Ð❭➂❼➒❸➘ ✗✖✕ ✁☎➻✵➁❳➂❤➐❁➄✍➹✰➅✚➉❛➋❫➚♠➃❸➎❤➃❸➌✚Ð▼➅❁➉❛➈❄➎❼➃❸➊✍➏❛➈❄➎Ý➄❛➈❄➉❛➚♠➃❸➊✍➈✏➆★➒✇➒❸➋✥➪❳➄✯➏❛➂✚➈ ➌✚➈✍➏❏➪✖➋♠➉❛✃✏➏❛➋✵➋♠➅❩➈❄➉r➆❫➏❛➈✵➪✖➃❸➏❛➂✞➆✯➂❤➃❸Ð♠➂✚➈✍➉✹➎✚➈✍Ð♠➉❛➈✍➈✂➋❫➑✿➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈✵➏❛➉r➆✍Ó☎➊ ➏❛➂➇➆❫➌⑧➪✖➋♠➐❤➒❸➎❞➶❩➈✏➆★➒✇➒❸➋✥➪✖➈✍➎❞➶❤➘⑧➆✪➅❁➐✚➉❛➈☎Ð♠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎❞➄❛➈✍➉❛➚♠➃❸➊✍➈✪➋✥Þ☛➈✍➉➍➓ ➃❸➌✚Ð✪➏❛➂✚➈✂➄r➆❫➔✵➈✂➎❁➈r➒❷➆★➘❋➶➇➋♠➐✚➌❁➎✚➄✍➻✹➁❳➂❁➈❄➄❛➈✯➉❛➈✍➄❛➐❤➒❸➏❛➄✍➹✰➏❛➂❁➋❭➐❁Ð❭➂Ñ➪✖➋❤➈r➑❦➐❼➒✇➒❸➘ ➃❸➌✚➊✍➋❭➔✵➅❼➒❸➈❄➏❛➈♠➹❢➆✥➉❛➈✹➷❤➐❩➆✍➒✇➃❸➏r➆❫➏➍➃❸➚❭➈➡➒③➘✞➊✍➋❭➌❁➄➍➃❸➄❛➏❛➈✍➌❤➏Ò➪✖➃❸➏❛➂❋➋♠➐✚➉✠➆✥➌❩➆✍➒❸➘❤➄➍➃❸➄✍➻ ➼✞➈☎➆❫➉❛➈☎➊✍➐✚➉❛➉❛➈✍➌❤➏➍➒❸➘⑧➆❫➏❛➏❛➈✍➔✵➅✚➏➍➃❸➌✚ÐÑ➏❛➋✏➔✵➋♠➉❛➈✯➑s➐❼➒✇➒❸➘▼➚❭➆✍➒✇➃❸➎❩➆✥➏❛➈✪➋♠➐✚➉ ➎✚➈✍➄➍➃❸Ð❭➌❑➏❛➂❁➉❛➋❭➐❁Ð❭➂✸➄➍➃❸➔✂➐❼➒❷➆✥➏➍➃❸➋♠➌✻➹➮➆❫➌✚➎✸➪✖➈✏➂❁➋❭➅❩➈❋➏❛➋▼➉❛➈✍➅➇➋♠➉❛➏✵➋❭➌✸➋♠➐✚➉ ➅✚➉❛➋♠Ð❭➉❛➈✍➄❛➄✳➃❸➌✏➆➴➄❛➐✚➶✚➄❛➈✍➷❤➐❁➈❄➌❤➏➴➅✚➐✚➶❼➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✻➻Ò➸❳➋♠➏❛➈✖➏❛➂➇➆❫➏✿➔✂➐❁➊r➂☎➋❫➑☛➏❛➂✚➈ ➊r➂➇➆★➒✇➒❸➈✍➌✚Ð♠➈✵➂✚➈✍➉❛➈✠➃❸➄❳➎✚➈✍➏❛➈✍➉❛➔✯➃❸➌❤➃❸➌❁Ð✏➂❁➋✥➪✫➏❛➋✵➈✍➚❭➆★➒❸➐➇➆❫➏❛➈✂➋♠➐✚➉❳➅❁➉❛➋❭➅❩➋♠➄r➆✍➒❦➻ ➁❳➂✚➈✍➉❛➈➴➃❸➄❳➌✚➋✯➪✖➃❸➎✚➈➡➒❸➘✏➆❫➊✍➊❄➈✍➅❁➏❛➈❄➎✪➄❛➈✍➏✖➋✥➑❢➶❩➈❄➌❁➊r➂❁➔➾➆❫➉❛✃❤➄✖➑❦➋♠➉✦➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈ ➒❸➋❤➆✥➎❁➄✍➹P➆❫➌✚➎⑤➔✂➐❁➊r➂❑➋❫➑✲➏❛➂✚➈Ñ➌✚➋✥➚♠➈➡➒❸➏❏➘❑➋❫➑✲➋♠➐✚➉✪➐✚➌❼➃✇➽➇➈✍➎✛➄❛➊r➂❁➈❄➎❁➐❤➒✇➃❸➌❁Ð ➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔ ➃❸➄✂➋♠➐✚➉✯➅✚➉❛➋✥➚♠➃❸➄➍➃❸➋♠➌❞➑❦➋♠➉✹➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❞➄❛➈✍➉❛➚♠➃❸➊✍➈♠➹✿➪❳➂❤➃❸➊➡➂▼➊★➆❫➌ ➋♠➌❤➒❸➘☎➶❩➈✠➔✲➈✥➆❫➌❤➃❸➌❁Ð✥➑❦➐❤➒✇➒③➘❋➏❛➈✍➄❛➏❛➈❄➎➾➃❸➌✏➆✹➎❁➘❤➌➇➆❫➔✂➃❸➊✯➈✍➌❤➚♠➃❸➉❛➋❭➌❁➔✲➈✍➌❤➏✠➪✖➃❸➏❛➂ ➆❫➎➇➆❫➅✚➏➍➃❸➚♠➈✲➊➡➒✇➃❸➈✍➌❤➏❛➄✍➻ ➼Ñ➈☎➂➇➆★➚❭➈☎➌❁➋✥➪✽➊✍➋♠➔✲➅❼➒❸➈✍➏❛➈❄➎Ý➏❛➂✚➈✵➽❩➉❛➄❛➏✹➅➇➆❫➉❛➏❛➄✂➋❫➑Ô➋♠➐✚➉✵➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊➡➓ ➏❛➐✚➉❛➈♠➻▼➼✞➈✪➂➇➆✍➚♠➈✪➎✚➈✍➄❛➊✍➉➍➃❸➶➇➈✍➎✫➆✪➔✵➋❤➎✚➈➡➒❳➑❦➋♠➉✂➏❛➂❁➈☎➒❸➋✥➪Ô➓❦➒❸➈✍➚❭➈➡➒✹➅➇➆❫➊r✃♠➈❄➏ ➑❦➋❭➉❛➪❳➆❫➉❛➎❤➃❸➌❁Ð✲➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔✪➹❩➪❳➂❤➃❸➊➡➂✵➃❸➄❳➆✠➄❛➂➇➆❫➉➍➃❸➌✚Ð✵➎❤➃❸➄❛➊➡➃❸➅❤➒✇➃❸➌❩➈P➃❸➌✚➄➍➃❸➎✚➈✂➆❫➌ ➃❸➄❛➋❫➒❷➆✥➏➍➃❸➋♠➌❑➎❤➃❸➄❛➊➡➃❸➅❤➒✇➃③➌✚➈♠➹Ò➆❫➌✚➎❞➪✖➈☎➂❩➆✍➚❭➈✪➅✚➉❛➋❫➚♠➃❸➎✚➈✍➎ ➆✪➅❩➆✥➉❛➏➍➃❸➊✍➐❼➒❷➆✥➉☎➈✍➬♠➓ ➆❫➔✲➅❼➒❸➈✂➋❫➑✿➄❛➐✚➊➡➂✞➆✂➄❛➊➡➂✚➈✍➔✵➈♠➹☛➪❳➂❤➃❸➊➡➂❋➅✚➉❛➋✥➚♠➃❸➎❁➈❄➄✹➶❩➋♠➏❛➂❋➋✥➑❢➋♠➐✚➉✠➄❛➈❄➉❛➚♠➃❸➊✍➈ ➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈✍➌❤➏❢➔✲➋❤➎❁➈➡➒❸➄✍➹äÐ♠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎☎➆❫➌✚➎✹➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎✧➻✦➁❳➂❁➈✿➄❛➊➡➂✚➈✍➔✵➈ ➅✚➉❛➋❫➚♠➃❸➎✚➈✍➄✿➄❛➈✍➚♠➈❄➉r➆★➒➇➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✲➄❛➈✍➉❛➚♠➃❸➊✍➈❳➊➡➒❷➆❫➄❛➄❛➈❄➄✳➪✖➃❸➏❛➂✂➎❼➃✇Þ☛➈❄➉❛➈✍➌❤➏✳➎✚➈➡➒❷➆✍➘ ➶❩➋❭➐❁➌✚➎❁➄✍➹➇➆✥➌❁➎✵➐✚➄❛➈✍➄✿➆✠➅➇➆❫➉❛➏➍➃❸➊✍➐❤➒❷➆❫➉✦➏❛➈✍➊➡➂✚➌❤➃❸➷❤➐❁➈✂➭ ✛✻➯✢✛✽✫ ✰ ➺☛➏❛➋✠➅✚➉❛➋✥➚♠➃❸➎❁➈ ➒❸➋✥➪✯✲➍➃❸➏❛➏❛➈✍➉✍➹❙➆❫➌✚➎❋➏❛➋✵➅✚➉❛➋❫➚♠➃❸➎✚➈☎➆❂✲❏➃❸➏❛➏❛➈✍➉❳➶❩➋♠➐✚➌❁➎✞➏❛➂➇➆❫➏❳➎❁➋❤➈❄➄✠➌✚➋♠➏✠➚❭➆✥➉❛➘ ➄❛➏❛➉❛➋♠➌✚Ð❫➒❸➘❋➪✖➃❸➏❛➂✏➏❛➂❁➈✹➌❤➐✚➔✯➶➇➈✍➉✠➋✥➑✿➂❁➋❭➅❁➄❳➃❸➌✪➏❛➂✚➈✯➅➇➆❫➏❛➂✚➄✍➻ ✑✏❉✺❤❾✓✒♠Ü✇➀✔✺✪Ö❸Ø❤❽✢✺❤❾✽❏❿❤➀✲✺ ☎✠➄➾➆ ➅➇➆❫➉❛➏✵➋✥➑✹➏❛➂✚➈✪➎✚➈➡➽➇➌❼➃❸➏➍➃❸➋❭➌⑤➋✥➑✠➏❛➂✚➈✪➐✚➌❤➃✇➽❩➈✍➎❑➄❛➊r➂✚➈✍➎❁➐❤➒✇➃❸➌✚Ð❑➆★➒❸Ð❭➋❫➓ ➉➍➃❸➏❛➂✚➔✪➹✖➪✖➈☎➂➇➆✍➚♠➈❋➆✍➒❸➄❛➋Ñ➎✚➈➡➽➇➌❁➈❄➎✸➋♠➐✚➉✵➄❛➈✍➉❛➚♠➃❸➊❄➈✆➃❸➌❤➏❛➈✍➉➍➑❏➆✥➊✍➈♠➻⑧➯➍➌▼➑❏➆❫➊❄➏✍➹ ➏❛➂✚➈✍➉❛➈☎➆❫➉❛➈☎➏❏➪✖➋✵➑❦➋♠➉❛➔✵➄✹➑❦➋❭➉✠➏❛➂✚➈✆➄❛➈❄➉❛➚♠➃❸➊✍➈✵➃❸➌❤➏❛➈✍➉➍➑❏➆❫➊❄➈♠➹❢➋♠➌✚➈✵➑❦➋♠➉✹Ð❭➐❩➆✥➉➍➓ ➆❫➌❤➏❛➈❄➈✍➎❋➄❛➈✍➉❛➚♠➃❸➊✍➈✲➆❫➌✚➎✞➆❫➌✚➋♠➏❛➂✚➈✍➉✖➑❦➋♠➉❳➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻ ✛❁➋♠➉✰Ð♠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➎✵➄❛➈✍➉❛➚♠➃❸➊✍➈❭➹♠➏❛➂✚➈✳➃❸➌❤➏❛➈✍➉➍➑❏➆✥➊✍➈✳➃❸➄✰➄➍➃❸➔✵➅❼➒❸➈❭❒✳➏❛➂✚➈✿➄❛➋♠➐✚➉❛➊✍➈ ➋♠➌❤➒❸➘❑➌✚➈✍➈✍➎✚➄✪➏❛➋❞➄❛➅❩➈✍➊➡➃✇➑s➘❑➏❛➂❁➈✞➌❁➈❄➈✍➎❁➈❄➎❑➊➡➒❸➋❤➊➡✃❑➉r➆✥➏❛➈✢✠ ✹ ➹Ò➏❛➂✚➈✍➌❑➏❛➂❁➈ ➌✚➈✍➏➍➪✖➋❭➉❛✃❞Ð♠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➄✪➏❛➂❤➃❸➄✪➉r➆✥➏❛➈♠➻✫➁➴➂✚➈❋➄❛➋♠➐✚➉❛➊✍➈✞➐❁➄❛➈❄➄➾➃❸➏❛➄✏✃❤➌❁➋✥➪❳➌ ➚❭➆★➒❸➐✚➈✵➑❦➋♠➉✸☛ ✹ ➭✒✠ ✹ ➺✠➏❛➋✪➊✍➋❭➔✵➅❁➐✚➏❛➈✵➃❸➏❛➄✯➪✦➋♠➉❛➄❛➏✹➊✥➆✥➄❛➈✵➷❤➐❁➈❄➐❁➈➡➃❸➌✚Ð✞➎❁➈➡➒❷➆✍➘♠➻ ➯◗➑✂➏❛➂✚➈❋➎✚➈➡➒❷➆✍➘✸➃❸➄✪➐✚➌❁➄❛➐❤➃❸➏r➆❫➶❤➒❸➈♠➹✹➃❸➏✪➔✂➐❁➄❛➏✏➉❛➈✍➷❤➐✚➈✍➄❛➏❋➆Ñ➂❤➃❸Ð♠➂✚➈✍➉❋➊➡➒❸➋❤➊r✃ ➉r➆❫➏❛➈✡✠ ✹ ➻⑧➁❳➂❁➈✏➌❁➈✍➏❏➪✖➋♠➉❛✃❞➎✚➋❤➈✍➄✵➌✚➋▼➊✍➋❭➌❼➑❦➋❭➉❛➔☎➆❫➌✚➊✍➈❋➊r➂❁➈❄➊➡✃▼➋❭➌❑➆✥➌❤➘ Ð♠➐➇➆❫➉r➆✥➌❤➏❛➈✍➈✍➎☎➄❛➈✍➉❛➚♠➃❸➊✍➈❳➱❩➋✥➪✹➹❤➶❩➈❄➊✥➆❫➐✚➄❛➈✠➏❛➂✚➈✖➱❩➋✥➪❞➎❁➋❤➈❄➄✖➌❁➋❭➏✳➔➾➆❫✃♠➈Ò➆✥➌❤➘ ➏❛➉r➆★Ó➾➊✹➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌▼➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈❄➌❤➏✠➏❛➋✵➏❛➂✚➈✹➌❁➈❄➏➍➪✖➋❭➉❛✃❩➻ ✛❁➋♠➉Ò➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈❭➹☛➏❛➂❁➈✂➄❛➈✍➉❛➚♠➃❸➊✍➈✯➃❸➌❤➏❛➈❄➉➍➑❏➆❫➊✍➈✂➔✂➐❁➄❛➏Ò➊➡➂➇➆❫➉r➆✥➊➡➓ ➏❛➈✍➉➍➃❸à✍➈✹➶➇➋♠➏❛➂✪➏❛➂✚➈✠➏❛➉r➆✍Ó☎➊✹➆✥➌❁➎✏➏❛➂❁➈✹➄❛➈❄➉❛➚♠➃❸➊✍➈♠➻★✛❁➋♠➉✦➏❛➂❁➈✹➊r➂❩➆✥➉r➆❫➊✍➏❛➈❄➉➍➃❸à✥➆★➓ ➏➍➃❸➋♠➌☎➋✥➑❢➏❛➂✚➈➴➏❛➉r➆✍Ó☎➊✠➪✦➈➴➂➇➆✍➚♠➈Ò➏❛➂❁➈Ò➄❛➋♠➐✚➉❛➊✍➈✠➎✚➈✍➊➡➒❷➆✥➉❛➈✹➏❛➂❁➈❳➅➇➆❫➉r➆❫➔✲➈✍➏❛➈✍➉❛➄ ➭✒✠✔✓✄☛➡➺❳➋❫➑✦➏❛➂❁➈✂➏❛➋♠✃❭➈✍➌✞➶❁➐✚➊➡✃❭➈✍➏✠➏❛➉r➆✍Ó☎➊✹➽❁➒❸➏❛➈✍➉✹➏❛➋☎➪❳➂❤➃❸➊➡➂❋➃❸➏✠➊➡➒❷➆✍➃❸➔✵➄✹➃❸➏❛➄ ➏❛➉r➆★Ó➾➊✆➪✖➃✇➒✇➒❳➊❄➋♠➌❤➑❦➋♠➉❛➔✪➻ ➸➴➋♠➏❛➈✲➏❛➂❩➆✥➏P➃❸➌❞➏❛➂✚➈✆Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎Ý➊★➆❫➄❛➈☎➏❛➂❁➈ ➊➡➒✇➃❸➈✍➌❤➏Ò➎❼➃❸➎✏➌❁➋❭➏✖➌✚➈✍➈✍➎✪➏❛➋✹➃❸➌❤➑❦➋♠➉❛➔✾➏❛➂✚➈✠➌✚➈✍➏➍➪✦➋♠➉❛✃✵➋✥➑✻➃❸➏❛➄✖➶✚➐❁➊r✃♠➈❄➏➴➄➍➃❸à❄➈✂✁ ☛✍➻❋➲❤➈✍➅➇➆❫➉r➆❫➏❛➈r➒❸➘♠➹✿➏❛➂❁➈☎➄❛➋♠➐✚➉❛➊✍➈☎➔✂➐❁➄❛➏✂➉❛➈✍➷❤➐✚➈✍➄❛➏✯➏❛➂✚➈☎➌❁➈❄➈✍➎❁➈❄➎▼➄❛➈✍➉❛➚♠➃❸➊✍➈♠➻ ➁❳➂❼➃❸➄❳➃❸➌❤➚♠➋✥➒❸➚♠➈✍➄✂➄❛➈➡➒❸➈✍➊✍➏➍➃❸➌✚Ð✞➆✵➄❛➐❼➃❸➏r➆✥➶❼➒❸➈☎➎✚➈➡➒❷➆✍➘ ✟ ➆✥➌❁➎⑧➆✯➏r➆✥➉❛Ð♠➈✍➏❳➒❸➋♠➄❛➄ ➉r➆❫➏❛➈☎✄✛➏❛➂✚➈✪➆✥➅❁➅❤➒✇➃❸➊✥➆✥➏➍➃❸➋♠➌✫➊✥➆❫➌▼➏❛➋✥➒❸➈✍➉r➆❫➏❛➈❭➻▼➁❳➂❁➈☎➌✚➈✍➏➍➪✦➋♠➉❛✃▼➪✖➃✇➒✇➒Ò➐❁➄❛➈ ➏❛➂✚➈✍➄❛➈✯➌❤➐✚➔✂➶❩➈✍➉❛➄✠➏❛➋➾➆❫➄❛➄➍➃❸Ð♠➌❋➏❛➂✚➈✂➄❛➋♠➐✚➉❛➊✍➈✹➏❛➋✏➆❫➌❋➆✥Ð♠Ð❭➉❛➈✍Ð❤➆✥➏❛➈✯➊➡➒❷➆✥➄❛➄✹➆✥➏ ➈✥➆✥➊➡➂▼➄❛➪✖➃❸➏❛➊r➂▼➑❦➋♠➉✹➄❛➂➇➆❫➉➍➃❸➌✚ÐÑ➅✚➐✚➉❛➅❩➋♠➄❛➈✍➄✍➻✞➁❳➂❤➐✚➄✍➹✻➑❦➋♠➉✂➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎❞➄❛➈✍➉➍➓ ➚♠➃❸➊✍➈♠➹ä➏❛➂❁➈✿➅❩➆✥➉r➆❫➔✵➈❄➏❛➈✍➉❛➄❢➋❫➑➇➏❛➂❁➈✦➄❛➈✍➉❛➚♠➃❸➊✍➈✳➃❸➌❤➏❛➈✍➉➍➑◗➆❫➊✍➈❳➆✥➉❛➈✖➏❛➂✚➈ ➽✚➒❸➏❛➈✍➉❢➉r➆❫➏❛➈ ➆❫➌✚➎❋➄➍➃❸à✍➈✵➭✒✠✔✓✙☛➡➺❳➆❫➌✚➎❋➏❛➂✚➈✹➎❁➈➡➒❷➆✍➘ ➆❫➌✚➎✪➒❸➋♠➄❛➄✠➊r➂➇➆❫➉r➆❫➊❄➏❛➈✍➉➍➃❸➄❛➏➍➃❸➊✍➄✵➭✟ ✓✆✄✿➺➡➻ ➁❢➋✵➅❁➉❛➋✥➚♠➃❸➎✚➈☎➅❁➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✞➄❛➈✍➉❛➚♠➃❸➊✍➈❭➹✻➏❛➂✚➈✯➌✚➈✍➏❏➪✖➋♠➉❛✃✏➔✯➐✚➄❛➏✹➆✍➒❸➄❛➋✪➈✍➌❤➓ ➑❦➋❭➉❛➊✍➈✠➏❛➂✚➈✠➏❛➉r➆✍Ó☎➊✠➊✍➋❭➔✵➔✯➃❸➏❛➔✵➈❄➌❤➏❛➄➴➔➾➆❫➎✚➈✠➶❤➘☎➏❛➂❁➈Ò➊➡➒✇➃❸➈✍➌❤➏❛➄✍➻ ✎➌❤➑❦➋♠➉❛➊❄➈➡➓ ➔✵➈✍➌❤➏✠➃❸➄✹➊★➆❫➉❛➉➍➃❸➈✍➎▼➋❭➐❁➏✂➆✥➄✠➑❦➋❫➒✇➒❸➋✥➪❳➄✍➻ ✎✿➆❫➊r➂Ñ➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎▼➄❛➈✍➉❛➚♠➃❸➊❄➈➾➱➇➋✥➪ ➃❸➄✂➊➡➂✚➈✍➊r✃♠➈✍➎⑧➆❫➏✹➏❛➂✚➈✵➈✍➎✚Ð♠➈☎➋❫➑✦➏❛➂✚➈✵➌❁➈❄➏➍➪✖➋❭➉❛✃▼➭③➃❦➻ ➈♠➻❸➹✰➏❛➂❁➈✯➽➇➉❛➄❛➏✂➄❛➪Ô➃❸➏❛➊r➂ ✁ ð✻ì✟✘ è✖é➡è❦è❦ò✟✚✖ç❛è✿í◗ì❭é➡í✿í❦ì✥ç➴è❦ñ✦✘ í❦ë➍ì✥ç❛è✦é❀✩ñ✰é❛÷✍è✳ì❭é✔✙➡ç❳è❦ò✏❭Òë✔✘ ç❛ÿ✍í✖ö✥ò☛❁✚ç❛ê✆✘ ÿ✥û✒▼ ❬✍ÿ✥ôrñ✽✩ ç❛õ✥û✍ç❳ô★ í❦ì✥ç❳ö❫ò✥ë❬➡ç❛í✿è✆✘ ✝❛ç✟✞ ✘ è✿ê◗ç✝✵❄ò✣✘ ê❦ç❛õ ★ô✍ê✿í◗ì✥ç❳è❦ñ✽✘ í◗ë❏ì✯í❦ô ✚❳é➡ÿ❭é➡û✍ç í❦ì✥ç✳ö✥ò☛❁✚ç❛ê❦è✼✘ ÿ✹ëré➡è❦ç❛è☛ô★ ô❀✙rç❛ê✥✲❤ôrñ ✪
he traffic passes through) for conformance to its declared ely as an illustration of the considerations involved in conforming packets are dropped or ng an admission control policy. It is clear that the gged. This conformance check provides the necessary is y of our proposal rests on our ability to formulat lat ion that is mandatory for entering a shared worl an admIs ntr rhich will make the predicted hat initial check, conformance is never enforced at later service class sufficient ly reliable specifying and validating such an admission control policy is the focus of our current to the scheduling policies and load dynamics of the net work and not the generation be havior of the source. We have the following additional In the case of the predicted service, specifying the token admission control policies. It is not clear how precise such bucket traffic filter also permits the network to est imate if it an algorithm needs to be. If there is enough bandwidth to an carry the new source at the requested rate and burstiness d if only a smal and still meet the service targets for this, and all of the traffic needs the most demanding of the predicted service, xisting, flows. This is the function of the last part of the then a rough estimate may be adequate. In addition, we architecture, the flow admission control computation are offering a general method which involves measuring the behavior of the existing real-time traffic, rather than usin 9 Admission Control the traffic model specified in the service interface, in decid ing whe her to admit new traffic. We use the worst-case we stated earlier that we would not address the ne traffic model only for the new source, which we cannot oth on process for the esta blishment of service comm we must at least address the conditions under which will be able to measure the aggregate traffic with the ne a net work accepts or denies a request for service, wit hout flow and base further admission decisions on the most necessarily specifying the exact d ynamics of that exchange measurement. This approach is important for t wo reasons here are two criteria to apply when deciding whethe Firs or not to admit additional flows into the network. The first limits, this will give a better characterization and better admission control criterion is that we should reserve no m link utilization. Second, it matches what the clients them than 90u of the bandwidth for real-time traffic, there by let ing the datagram traffic have access to at least xou of the observed network traffic. Having the kt while the numerical value, doU, of this quot a is com ssess the traffic in similar ways is likely to better letely ad hoc and experience may suggest other values ar produce reasonable behavior more effective. we do believe that it is crucial to have such a quota. This quota ensures that the datagram service 1G 9 ther service o ualities ains operational at all times thaving the datagram traffic filep etet i sibleodearanda oy tneg paetragds o trans wrt Droresede a the certex t oeresa-vime services e e we wish ayers. In addition, the datagram quota ensures that there to mention them, alt hough we do not discuss exactly wish enough spare capacity to accommodate sizable fluct ua ns in the guaranteed and predicted service traffic. The One goal is that if overload causes some of the packets ond admission cont rol criterion is that we want to from a sou miss their deadline should sure that the addition of a flow does not increase able to se parate its packe ts into different classes, to control We now give an example, albeit superficial, which packets get dropped. This idea can be incorporated how one into our scheme by creating several ty classes with tI might make these criteria specific. Let i denote some mea- same target Di. Packets tagged as less important go into sure of the utilization on a link due to realtime traffic(in the lower priority class, where they will arrive just behind ymbol denotes measured quantities), let g tant packets, but with higher le measure of the delay of the traffic in the classes with larger D obvious that use of priorit nd let u denote the link speed. In this example admission here can create a range of policies. control criterion, a flow promising to conform to a token Another proposed service is that packets that are suffi bucket traffic filter (r, b) can be admitted to priority level i iently late should be discarded internally, rather than be if()r wi<9u, and (2 )b<(D; -)(u-i-r)for each ing delive red, since in delivering them the network may use than or equal in priority to level. Th first condition guarantees that there is at least xou of the delay of su bsequent packets. The offset carried in the packe link left over for datagram traffic. The second condition is a in the FIFOw scheme provides precisely the needed infor- uristic designed to ensure that the delays will not violate mation: if a packet accumulates a very large jitter offset, it the bounds di once the new now is admitted even if the been pro new flow displays worst-case behavior. The key to making posed else where ([2rE but we observe that it fits naturally ict ive service commit ments relia ble is to choose ap to the fifo. scheme ropriately conservative measures for i and gi tthese should A third service is that packets should be buffered in th but consistent ly conservative estimate twork if they might ot herwise arrive early(before the play at) so that the lot provide the bi Bnowing how conservative to make the i and g may involve ng or estimate the delay. We are not convinced that historical knowledge of the size of fluctuations in net work this service is useful in general. With current memory costs, traffic and delay on various link ve, And while it might see cly sketchy, and nice for the network to relieve the destination equipment this particular impleme from the need to estimate the delay, it d function in a dynam We offer it
➏❛➂✚➈☎➏❛➉r➆★Ó☎➊☎➅➇➆❫➄❛➄❛➈✍➄✂➏❛➂✚➉❛➋♠➐✚Ð♠➂➇➺P➑❦➋♠➉✂➊✍➋❭➌❼➑❦➋❭➉❛➔☎➆❫➌✚➊✍➈✏➏❛➋✆➃❸➏❛➄✵➎✚➈✍➊➡➒❷➆✥➉❛➈✍➎ ➏❛➋❭✃♠➈✍➌⑧➶❁➐✚➊➡✃❭➈✍➏✯➽✚➒❸➏❛➈✍➉☛✰✖➌✚➋♠➌✚➊✍➋❭➌❼➑❦➋❭➉❛➔✯➃❸➌✚Ð▼➅❩➆✥➊➡✃❭➈✍➏❛➄✲➆❫➉❛➈☎➎❁➉❛➋❭➅❁➅➇➈✍➎❞➋♠➉ ➏r➆✥Ð♠Ð❭➈✍➎✧➻✦➁❳➂❼➃❸➄❳➊✍➋❭➌❼➑❦➋❭➉❛➔☎➆❫➌✚➊✍➈✂➊➡➂✚➈✍➊r✃☎➅❁➉❛➋✥➚♠➃❸➎✚➈✍➄✠➏❛➂✚➈✹➌❁➈❄➊✍➈✍➄❛➄r➆❫➉❛➘☎➃❸➄❛➋❫➓ ➒❷➆✥➏➍➃❸➋♠➌✞➏❛➂❩➆✥➏✖➃❸➄✠➔☎➆✥➌❁➎➇➆❫➏❛➋❭➉❛➘✆➑❦➋❭➉➴➈❄➌❤➏❛➈✍➉➍➃❸➌❁Ð ➆✯➄❛➂➇➆❫➉❛➈✍➎❋➪✖➋❭➉➍➒❸➎✧➻★☎➴➑❦➏❛➈❄➉ ➏❛➂➇➆❫➏☎➃❸➌❼➃❸➏➍➃❷➆✍➒☎➊r➂❁➈✍➊r✃❩➹Ò➊✍➋♠➌❤➑❦➋♠➉❛➔➾➆❫➌✚➊✍➈Ñ➃❸➄✪➌✚➈✍➚❭➈✍➉✪➈✍➌❤➑❦➋❭➉❛➊✍➈✍➎✷➆❫➏✆➒❷➆✥➏❛➈✍➉ ➄❛➪✖➃❸➏❛➊r➂❁➈❄➄☛✰❳➏❛➂❼➃❸➄✯➃❸➄☎➶❩➈❄➊✥➆❫➐✚➄❛➈❋➆✥➌❤➘Ñ➒❷➆❫➏❛➈✍➉☎➚♠➃❸➋✥➒❷➆❫➏➍➃❸➋♠➌❑➪✦➋♠➐❤➒❸➎✸➶❩➈✏➎❁➐✚➈ ➏❛➋✂➏❛➂✚➈✠➄❛➊r➂❁➈✍➎✚➐❤➒✇➃❸➌❁Ð✏➅❩➋❫➒✇➃❸➊➡➃❸➈❄➄✹➆❫➌✚➎✆➒❸➋ä➆❫➎✏➎❁➘❤➌➇➆❫➔✯➃❸➊❄➄➴➋✥➑❢➏❛➂✚➈✠➌✚➈✍➏➍➪✦➋♠➉❛✃ ➆✥➌❁➎❋➌✚➋♠➏❳➏❛➂✚➈✹Ð♠➈✍➌✚➈✍➉r➆✥➏➍➃❸➋♠➌✞➶❩➈✍➂➇➆★➚♠➃❸➋❭➉✹➋❫➑✿➏❛➂✚➈✹➄❛➋♠➐✚➉❛➊✍➈♠➻ ➯➍➌✏➏❛➂❁➈Ò➊✥➆❫➄❛➈✹➋✥➑❢➏❛➂✚➈✠➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❋➄❛➈✍➉❛➚♠➃❸➊✍➈❭➹❩➄❛➅❩➈✍➊r➃✇➑❦➘♠➃❸➌❁Ð✏➏❛➂❁➈✹➏❛➋❭✃♠➈✍➌ ➶✚➐✚➊➡✃❭➈✍➏✖➏❛➉r➆★Ó➾➊Ô➽✚➒❸➏❛➈✍➉❳➆✍➒❸➄❛➋✯➅➇➈✍➉❛➔✯➃❸➏❛➄✖➏❛➂✚➈❳➌❁➈✍➏❏➪✖➋♠➉❛✃✂➏❛➋✹➈✍➄❛➏➍➃❸➔☎➆✥➏❛➈➴➃✇➑☛➃❸➏ ➊★➆❫➌✹➊✥➆✥➉❛➉❛➘✠➏❛➂✚➈✳➌✚➈✍➪▼➄❛➋❭➐❁➉❛➊✍➈Ô➆❫➏☛➏❛➂❁➈✿➉❛➈✍➷❤➐✚➈✍➄❛➏❛➈✍➎✂➉r➆❫➏❛➈✦➆❫➌✚➎✹➶❁➐✚➉❛➄❛➏➍➃❸➌✚➈✍➄❛➄ ➆✥➌❁➎✫➄❛➏➍➃✇➒✇➒✂➔✵➈✍➈❄➏✪➏❛➂✚➈❋➄❛➈✍➉❛➚♠➃❸➊✍➈▼➏r➆✥➉❛Ð♠➈✍➏❛➄☎➑❦➋♠➉✏➏❛➂❼➃❸➄✍➹➮➆❫➌✚➎✛➆✍➒✇➒✂➋❫➑✂➏❛➂✚➈ ➈❄➬♠➃❸➄❛➏➍➃❸➌❁Ð✚➹✳➱➇➋❫➪❳➄✍➻✲➁➴➂❤➃❸➄✹➃❸➄✹➏❛➂❁➈✵➑❦➐✚➌❁➊❄➏➍➃❸➋♠➌▼➋✥➑❳➏❛➂❁➈✂➒❷➆❫➄❛➏✂➅❩➆✥➉❛➏✹➋❫➑❳➏❛➂✚➈ ➆✥➉❛➊➡➂❤➃❸➏❛➈✍➊✍➏❛➐✚➉❛➈♠➹☛➏❛➂✚➈✠➱❩➋✥➪✷➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌⑧➊✍➋♠➌❤➏❛➉❛➋✥➒❢➊✍➋❭➔✵➅❁➐✚➏r➆❫➏➍➃❸➋❭➌✧➻ ×❹✳Ú❉❆✂Ü✇❻➡❻rÜ✇Ù❩Ø✆❊❢Ù❩Ø❤❽r❾◗Ù✹❂ ➼→➂❼➃✇➒❸➈✏➪✖➈✵➄❛➏r➆✥➏❛➈✍➎Ñ➈★➆❫➉➍➒✇➃❸➈✍➉✵➏❛➂➇➆❫➏✹➪✦➈✵➪✖➋♠➐❤➒❸➎▼➌✚➋♠➏✂➆✥➎❁➎✚➉❛➈✍➄❛➄✂➏❛➂❁➈✵➌✚➈➡➓ Ð❭➋♠➏➍➃❷➆✥➏➍➃❸➋♠➌❑➅✚➉❛➋❤➊✍➈✍➄❛➄✯➑❦➋❭➉✵➏❛➂✚➈✆➈❄➄❛➏r➆❫➶❤➒✇➃❸➄❛➂❁➔✵➈❄➌❤➏❋➋❫➑Ò➄❛➈✍➉❛➚♠➃❸➊✍➈❋➊✍➋❭➔✵➔✯➃❸➏➍➓ ➔✲➈✍➌❤➏❛➄✍➹✚➪✖➈✠➔✂➐❁➄❛➏Ô➆❫➏✳➒❸➈✥➆✥➄❛➏Ò➆❫➎✚➎❁➉❛➈✍➄❛➄❳➏❛➂✚➈✹➊✍➋♠➌✚➎❤➃❸➏➍➃❸➋♠➌✚➄✠➐✚➌❁➎✚➈✍➉❳➪❳➂❼➃❸➊r➂ ➆✪➌✚➈✍➏❏➪✖➋♠➉❛✃❞➆✥➊✍➊✍➈✍➅✚➏❛➄✵➋❭➉✯➎✚➈✍➌❤➃❸➈✍➄✏➆✪➉❛➈✍➷❤➐❁➈❄➄❛➏✯➑❦➋♠➉✂➄❛➈✍➉❛➚♠➃❸➊✍➈❭➹✖➪✖➃❸➏❛➂❁➋❭➐❁➏ ➌✚➈✍➊✍➈❄➄❛➄r➆❫➉➍➃✇➒❸➘✞➄❛➅❩➈✍➊➡➃✇➑❦➘♠➃❸➌✚Ð✪➏❛➂✚➈✹➈✍➬✚➆✥➊✍➏❳➎❁➘❤➌➇➆❫➔✯➃❸➊❄➄✹➋❫➑✿➏❛➂➇➆❫➏❳➈✍➬❤➊r➂❩➆✥➌❁Ð❭➈♠➻ ➁❳➂❁➈❄➉❛➈✪➆✥➉❛➈☎➏➍➪✖➋✏➊✍➉➍➃❸➏❛➈✍➉➍➃❷➆❋➏❛➋✞➆❫➅✚➅❼➒❸➘▼➪❳➂✚➈✍➌▼➎✚➈✍➊➡➃❸➎❤➃❸➌❁Ð⑧➪➴➂✚➈✍➏❛➂✚➈✍➉ ➋❭➉❳➌❁➋❭➏➴➏❛➋✲➆❫➎✚➔✯➃❸➏➮➆❫➎✚➎❼➃❸➏➍➃❸➋❭➌❩➆✍➒✳➱➇➋✥➪➴➄✿➃❸➌❤➏❛➋✆➏❛➂✚➈✠➌✚➈✍➏❏➪✖➋♠➉❛✃❩➻✿➁❳➂❁➈✠➽➇➉❛➄❛➏ ➆✥➎❁➔✂➃❸➄❛➄➍➃❸➋♠➌☎➊✍➋♠➌❤➏❛➉❛➋✥➒❩➊❄➉➍➃❸➏❛➈✍➉➍➃❸➋♠➌✵➃❸➄❢➏❛➂➇➆❫➏❢➪✖➈✿➄❛➂❁➋❭➐❼➒❸➎☎➉❛➈✍➄❛➈❄➉❛➚♠➈✖➌✚➋✠➔✵➋❭➉❛➈ ➏❛➂➇➆❫➌✘✓✬✕ ✁✛➋❫➑✰➏❛➂✚➈✠➶➇➆❫➌✚➎❁➪✖➃❸➎✚➏❛➂✆➑❦➋❭➉✳➉❛➈✥➆✍➒✇➓❏➏➍➃❸➔✵➈✹➏❛➉r➆★Ó☎➊❭➹❤➏❛➂❁➈❄➉❛➈✍➶❤➘✯➒❸➈✍➏➍➓ ➏➍➃❸➌✚Ð☎➏❛➂❁➈✹➎➇➆❫➏r➆✥Ð♠➉r➆✥➔ ➏❛➉r➆✍Ó☎➊✯➂➇➆✍➚♠➈✂➆✥➊✍➊✍➈✍➄❛➄Ò➏❛➋☎➆❫➏✖➒❸➈✥➆✥➄❛➏☛✂✔✕✄✁✔➋❫➑✿➏❛➂✚➈ ➒✇➃❸➌✚✃ ✰✿➪❳➂❼➃✇➒❸➈☎➏❛➂✚➈✵➌❤➐❁➔✲➈✍➉➍➃❸➊✥➆★➒✦➚❭➆★➒❸➐✚➈♠➹❏✂✭✕✁✆➹✰➋❫➑✦➏❛➂❼➃❸➄✂➷❤➐✚➋♠➏r➆➾➃❸➄✹➊❄➋♠➔✯➓ ➅❤➒❸➈✍➏❛➈➡➒❸➘❞➝❤➞✜✦❩➫✥➤❳➆❫➌✚➎❋➈✍➬❤➅❩➈✍➉➍➃❸➈✍➌✚➊✍➈✵➔➾➆★➘✏➄❛➐❁Ð❭Ð♠➈✍➄❛➏Ò➋♠➏❛➂✚➈✍➉✠➚❭➆✍➒❸➐❁➈✍➄✂➆✥➉❛➈ ➔✲➋♠➉❛➈✂➈➡Þ☛➈✍➊✍➏➍➃❸➚❭➈♠➹✰➪✖➈✯➎✚➋☎➶❩➈➡➒✇➃❸➈✍➚❭➈☎➏❛➂❩➆✥➏➴➃❸➏✠➃❸➄✠➊❄➉❛➐❁➊➡➃❷➆✍➒✖➏❛➋☎➂➇➆✍➚♠➈✵➄❛➐✚➊➡➂ ➆☎➷❤➐❁➋❭➏r➆❼➻✲➁➴➂❤➃❸➄✹➷❤➐✚➋♠➏r➆✪➈✍➌✚➄❛➐❁➉❛➈❄➄✹➏❛➂❩➆✥➏✹➏❛➂❁➈✵➎➇➆❫➏r➆✥Ð♠➉r➆✥➔ ➄❛➈❄➉❛➚♠➃❸➊✍➈☎➉❛➈➡➓ ➔➾➆★➃❸➌✚➄✹➋♠➅❩➈✍➉r➆✥➏➍➃❸➋♠➌➇➆★➒Ò➆✥➏✹➆✍➒✇➒✿➏➍➃❸➔✵➈✍➄☛✰✰➂❩➆✍➚♠➃❸➌✚Ð❋➏❛➂❁➈✂➎❩➆✥➏r➆❫Ð❭➉r➆❫➔❅➏❛➉r➆✍Ó☎➊ ➊❄➋♠➔✵➅❤➒❸➈✍➏❛➈➡➒❸➘❋➄❛➂❤➐✚➏❳➋♠➐✚➏Ô➑s➋♠➉Ò➆❫➉❛➶❤➃❸➏❛➉r➆❫➉➍➃✇➒❸➘✪➒❸➋❭➌❁Ð☎➅❩➈✍➉➍➃❸➋❤➎✚➄✹➋❫➑❱➏➍➃❸➔✵➈✂➪Ô➃✇➒✇➒ ➒✇➃❸✃❭➈➡➒❸➘✸➅✚➐✚➏➾➃❸➔✲➅❩➋♠➄❛➄➍➃❸➶❤➒❸➈▼➎✚➈✍➔☎➆❫➌✚➎✚➄✆➋❭➌❞➏❛➂❁➈✏➎❩➆✥➏r➆❫Ð♠➉r➆✥➔ ➏❛➉r➆❫➌✚➄❛➅❩➋♠➉❛➏ ➒❷➆✍➘❭➈✍➉❛➄✍➻✂➯➍➌✞➆❫➎✚➎❤➃❸➏➍➃❸➋♠➌☛➹✳➏❛➂✚➈✂➎❩➆✥➏r➆❫Ð♠➉r➆✥➔❅➷❤➐✚➋♠➏r➆✆➈❄➌❁➄❛➐✚➉❛➈✍➄✂➏❛➂❩➆✥➏✠➏❛➂✚➈✍➉❛➈ ➃❸➄☎➈✍➌✚➋♠➐✚Ð♠➂❑➄❛➅➇➆❫➉❛➈❋➊★➆❫➅➇➆❫➊➡➃❸➏❏➘❑➏❛➋▼➆✥➊✍➊✍➋♠➔✲➔✵➋❤➎❩➆✥➏❛➈Ñ➄➍➃❸à★➆❫➶❤➒❸➈ ➱➇➐❁➊❄➏❛➐❩➆✍➓ ➏➍➃❸➋❭➌❁➄✵➃❸➌❞➏❛➂✚➈✆Ð❭➐❩➆✥➉r➆❫➌❤➏❛➈✍➈❄➎❑➆❫➌✚➎❞➅❁➉❛➈❄➎❼➃❸➊✍➏❛➈❄➎✸➄❛➈✍➉❛➚♠➃❸➊✍➈✏➏❛➉r➆★Ó☎➊❭➻Ñ➁❳➂✚➈ ➄❛➈❄➊✍➋♠➌✚➎✫➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌✫➊✍➋❭➌❤➏❛➉❛➋❫➒✹➊❄➉➍➃❸➏❛➈✍➉➍➃❸➋♠➌✸➃❸➄☎➏❛➂➇➆❫➏☎➪✖➈✪➪Ô➆❫➌❤➏☎➏❛➋Ñ➈❄➌❼➓ ➄❛➐✚➉❛➈✹➏❛➂➇➆❫➏✖➏❛➂✚➈✹➆✥➎❁➎❤➃❸➏➍➃❸➋♠➌❋➋✥➑✿➆✠➱❩➋✥➪❑➎✚➋❤➈✍➄❳➌❁➋❭➏✳➃❸➌❁➊❄➉❛➈✥➆❫➄❛➈✲➆❫➌❤➘✵➋✥➑❢➏❛➂✚➈ ➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎✞➎❁➈r➒❷➆★➘❤➄Ò➋✥➚♠➈✍➉❳➏❛➂✚➈✯➶❩➋❭➐❁➌✚➎✚➄✞✟✖ ➻ ➼✞➈✵➌❁➋✥➪✽Ð❫➃❸➚❭➈✏➆❫➌✞➈✍➬✚➆❫➔✵➅❤➒❸➈♠➹✦➆★➒❸➶➇➈➡➃❸➏✯➄❛➐✚➅❩➈✍➉➍➽➇➊➡➃❷➆✍➒❦➹✖➋❫➑Ô➂❁➋✥➪✽➋♠➌✚➈ ➔✂➃❸Ð♠➂❤➏✠➔➾➆❫✃♠➈✂➏❛➂✚➈✍➄❛➈✹➊✍➉➍➃❸➏❛➈✍➉➍➃❷➆☎➄❛➅❩➈✍➊➡➃✇➽➇➊♠➻ ✳✧➈❄➏✂✁✄ ➎✚➈✍➌✚➋♠➏❛➈✂➄❛➋♠➔✵➈✹➔✲➈✥➆★➓ ➄❛➐✚➉❛➈✵➋✥➑❳➏❛➂❁➈✲➐❁➏➍➃✇➒✇➃❸à✥➆✥➏➍➃❸➋♠➌→➋♠➌❞➆✵➒✇➃❸➌❁✃✞➎❁➐✚➈☎➏❛➋☎➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✏➏❛➉r➆★Ó☎➊☎➭③➃❸➌ Ð❭➈✍➌✚➈✍➉r➆★➒❦➹➇➏❛➂❁➈❳➂➇➆❫➏✿➄❛➘❤➔✯➶❩➋✥➒✰➎❁➈✍➌✚➋♠➏❛➈❄➄✖➔✵➈✥➆✥➄❛➐❁➉❛➈✍➎☎➷❤➐➇➆❫➌❤➏➍➃❸➏➍➃❸➈✍➄r➺➡➹❩➒❸➈❄➏ ✁☎ ✖ ➎✚➈✍➌✚➋♠➏❛➈✏➄❛➋♠➔✵➈✏➔✵➈✥➆❫➄❛➐✚➉❛➈✪➋✥➑Ò➏❛➂❁➈☎➎✚➈➡➒❷➆✍➘Ý➋✥➑Ò➏❛➂❁➈✏➏❛➉r➆★Ó☎➊✵➃❸➌❞➊➡➒❷➆❫➄❛➄ ✗➍➹ ➆✥➌❁➎✪➒❸➈❄➏✏✽Ñ➎✚➈✍➌✚➋♠➏❛➈✂➏❛➂❁➈✹➒✇➃❸➌❁✃❋➄❛➅➇➈✍➈✍➎☛➻❳➯➍➌✪➏❛➂❤➃❸➄✹➈✍➬✚➆✥➔✵➅❼➒❸➈➾➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌ ➊❄➋♠➌❤➏❛➉❛➋❫➒➮➊✍➉➍➃❸➏❛➈✍➉➍➃❸➋❭➌✧➹✂➆❋➱❩➋✥➪❅➅✚➉❛➋♠➔✯➃❸➄➍➃❸➌✚Ð✸➏❛➋⑧➊✍➋♠➌❤➑❦➋♠➉❛➔✒➏❛➋❞➆Ñ➏❛➋❭✃♠➈✍➌ ➶✚➐✚➊➡✃❭➈✍➏✠➏❛➉r➆✍Ó☎➊P➽✚➒❸➏❛➈✍➉✂➭✒✠✭✓✄☛r➺➴➊★➆❫➌❋➶❩➈✲➆❫➎✚➔✯➃❸➏❛➏❛➈✍➎✞➏❛➋✵➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘✪➒❸➈❄➚♠➈➡➒ ✗ ➃✇➑❳➭✢✂★➺✟✠ ✰ ✁✄✝✆✟✞ ✓ ✽❱➹➇➆✥➌❁➎✞➭ ✮♠➺✟☛ ✆ ➭✟✡✠ ✲ ✁☎☛✠✥➺➡➭✒✽ ✲ ✁✄ ✲ ✠✥➺✦➑s➋♠➉✖➈✥➆✥➊➡➂ ➊r➒❷➆❫➄❛➄✌☞✠➪❳➂❤➃❸➊➡➂✂➃❸➄ ➒❸➋✥➪✖➈✍➉❢➏❛➂➇➆❫➌✲➋♠➉❢➈✍➷❤➐➇➆★➒✚➃❸➌✵➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘✲➏❛➋➴➒❸➈✍➚❭➈➡➒ ✗➍➻❢➁❳➂✚➈ ➽➇➉❛➄❛➏✹➊✍➋❭➌❁➎❤➃❸➏➍➃❸➋♠➌→Ð♠➐➇➆❫➉r➆❫➌❤➏❛➈❄➈✍➄✯➏❛➂➇➆❫➏✹➏❛➂✚➈✍➉❛➈✵➃❸➄✂➆❫➏✠➒❸➈★➆❫➄❛➏✆✂✔✕✁ ➋❫➑❳➏❛➂✚➈ ➒✇➃❸➌✚✃✵➒❸➈➡➑❦➏✖➋✥➚❭➈✍➉ ➑s➋♠➉✿➎❩➆✥➏r➆❫Ð♠➉r➆✥➔✾➏❛➉r➆★Ó☎➊❭➻❢➁❳➂❁➈❳➄❛➈✍➊✍➋❭➌❁➎☎➊✍➋❭➌❁➎❤➃❸➏➍➃❸➋♠➌✪➃❸➄Ô➆ ➂✚➈✍➐✚➉➍➃❸➄❛➏➍➃❸➊✵➎✚➈✍➄➍➃❸Ð♠➌✚➈✍➎✞➏❛➋✵➈✍➌✚➄❛➐❁➉❛➈✂➏❛➂❩➆✥➏❳➏❛➂❁➈✹➎✚➈➡➒❷➆✍➘❤➄✹➪✖➃✇➒✇➒❢➌✚➋♠➏❳➚♠➃❸➋❫➒❷➆✥➏❛➈ ➏❛➂✚➈✪➶➇➋♠➐✚➌❁➎✚➄ ✟✡✠✏➋❭➌❁➊✍➈✏➏❛➂❁➈✏➌❁➈❄➪✔➱❩➋✥➪✔➃❸➄✏➆❫➎✚➔✯➃❸➏❛➏❛➈✍➎→➈✍➚♠➈✍➌❞➃✇➑✹➏❛➂✚➈ ➌✚➈✍➪✛➱➇➋❫➪✽➎❤➃❸➄❛➅❼➒❷➆✍➘❤➄✵➪✦➋♠➉❛➄❛➏➍➓❏➊✥➆❫➄❛➈✂➶❩➈✍➂➇➆✍➚♠➃❸➋♠➉✍➻☎➁❳➂✚➈✵✃♠➈✍➘❋➏❛➋✏➔☎➆❫✃♠➃❸➌✚Ð ➏❛➂✚➈✹➅✚➉❛➈✍➎❼➃❸➊❄➏➍➃❸➚♠➈✂➄❛➈✍➉❛➚♠➃❸➊✍➈✂➊✍➋♠➔✵➔✂➃❸➏❛➔✵➈✍➌❤➏❛➄❳➉❛➈➡➒✇➃❷➆❫➶❤➒❸➈✯➃❸➄❳➏❛➋✂➊➡➂✚➋❤➋♠➄❛➈✂➆✥➅❼➓ ➅✚➉❛➋♠➅✚➉➍➃❷➆❫➏❛➈r➒❸➘✆➊❄➋♠➌✚➄❛➈✍➉❛➚❭➆❫➏➍➃❸➚♠➈✹➔✲➈✥➆❫➄❛➐✚➉❛➈✍➄❢➑❦➋♠➉✍✁✄ ➆❫➌✚➎ ✁☎☛✠❏✰❤➏❛➂✚➈✍➄❛➈❳➄❛➂❁➋❭➐❼➒❸➎ ➌✚➋♠➏✍✲❛➐✚➄❛➏✳➶➇➈✠➆✍➚❭➈✍➉r➆❫Ð❭➈✍➄✳➶✚➐✚➏✳➊✍➋❭➌❁➄➍➃❸➄❛➏❛➈❄➌❤➏➍➒❸➘✆➊❄➋♠➌✚➄❛➈✍➉❛➚❭➆❫➏➍➃❸➚♠➈✹➈❄➄❛➏➍➃❸➔☎➆❫➏❛➈✍➄✍➻ ✎✠➌✚➋✥➪✖➃❸➌❁Ð✹➂✚➋✥➪Ý➊❄➋♠➌✚➄❛➈✍➉❛➚❭➆❫➏➍➃❸➚♠➈❳➏❛➋Ò➔☎➆❫✃❭➈✖➏❛➂❁➈✏✁✄ ➆✥➌❁➎ ✁☎✠ ➔➾➆★➘✠➃❸➌❤➚❭➋❫➒❸➚❭➈ ➂❤➃❸➄❛➏❛➋♠➉➍➃❸➊★➆★➒✂✃❤➌✚➋❫➪✖➒❸➈✍➎✚Ð♠➈❋➋✥➑Ò➏❛➂❁➈✏➄➍➃❸à✍➈❋➋❫➑❳➱➇➐❁➊✍➏❛➐➇➆❫➏➍➃❸➋❭➌❁➄☎➃❸➌✸➌✚➈✍➏➍➪✦➋♠➉❛✃ ➏❛➉r➆✍Ó☎➊✂➆❫➌✚➎❋➎✚➈➡➒❷➆✍➘✪➋♠➌✏➚❭➆❫➉➍➃❸➋♠➐✚➄✠➒✇➃❸➌✚✃❤➄✍➻ ➁❳➂❼➃❸➄✦➈✍➬✚➆❫➔✵➅❤➒❸➈✠➃❸➄❳➋✥➚♠➈✍➉➍➒❸➘✵➄❛✃❭➈✍➏❛➊r➂❤➘♠➹➇➆❫➌✚➎✆➪✦➈➴➂➇➆✍➚♠➈✹➘❭➈✍➏✳➏❛➋✂➄➍➃❸➔✂➐❼➓ ➒❷➆✥➏❛➈✯➏❛➋✂➄❛➈✍➈✹➂✚➋✥➪⑤➏❛➂❤➃❸➄❳➅❩➆✥➉❛➏➍➃❸➊✍➐❼➒❷➆✥➉✠➃❸➔✵➅❤➒❸➈✍➔✵➈✍➌❤➏r➆✥➏➍➃❸➋♠➌▼➋✥➑✦➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌ ➊❄➋♠➌❤➏❛➉❛➋❫➒Ô➪✖➋♠➐❤➒❸➎▼➑❦➐❁➌✚➊✍➏➍➃❸➋❭➌Ý➃❸➌→➆✪➎❁➘❤➌➇➆❫➔✂➃❸➊❋➌❁➈✍➏❏➪✖➋♠➉❛✃❩➻ ➼Ñ➈☎➋❫Þ❙➈✍➉✯➃❸➏ ➄❛➋❫➒❸➈r➒❸➘✛➆✥➄✞➆❫➌✸➃✇➒✇➒❸➐✚➄❛➏❛➉r➆❫➏➍➃❷➋♠➌✽➋❫➑☎➏❛➂✚➈✞➊✍➋♠➌✚➄➍➃❸➎❁➈❄➉r➆❫➏➍➃❸➋♠➌✚➄▼➃❸➌❤➚♠➋✥➒❸➚♠➈✍➎⑤➃❸➌ ➎✚➈✍➄➍➃❸Ð♠➌❤➃❸➌❁Ð→➆✥➌❞➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌❑➊❄➋♠➌❤➏❛➉❛➋❫➒Ô➅❩➋❫➒✇➃❸➊✍➘❭➻▼➯➍➏✹➃❸➄✵➊➡➒❸➈★➆❫➉✵➏❛➂➇➆❫➏✂➏❛➂❁➈ ➚♠➃❷➆❫➶❤➃✇➒✇➃❷➏➍➘❑➋❫➑➮➋♠➐✚➉☎➅❁➉❛➋❭➅❩➋♠➄r➆✍➒✂➉❛➈✍➄❛➏❛➄✵➋♠➌→➋♠➐✚➉✏➆❫➶❤➃✇➒✇➃❸➏➍➘❑➏❛➋❋➑❦➋♠➉❛➔✂➐❼➒❷➆✥➏❛➈ ➆❫➌①➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌→➊✍➋♠➌❤➏❛➉❛➋✥➒✖➅❩➋✥➒✇➃❸➊✍➘▼➪❳➂❤➃❸➊➡➂❞➪✖➃✇➒✇➒✖➔➾➆❫✃♠➈☎➏❛➂✚➈✵➅✚➉❛➈✍➎❼➃❸➊❄➏❛➈✍➎ ➄❛➈✍➉❛➚♠➃❸➊✍➈❋➊r➒❷➆❫➄❛➄☎➄❛➐❼Ó➾➊➡➃❸➈✍➌❤➏➍➒❸➘✸➉❛➈r➒✇➃❷➆❫➶❤➒❸➈✒✰✹➄❛➅❩➈✍➊r➃✇➑❦➘♠➃❸➌❁Ð→➆✥➌❁➎→➚♠➆✍➒✇➃❸➎❩➆✥➏➍➃❸➌❁Ð ➄❛➐✚➊➡➂✞➆✥➌❋➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋❭➌▼➊✍➋♠➌❤➏❛➉❛➋✥➒❢➅❩➋✥➒✇➃❸➊✍➘✪➃❸➄✠➏❛➂✚➈✠➑❦➋❤➊✍➐✚➄❳➋❫➑✿➋❭➐❁➉Ò➊✍➐❁➉❛➉❛➈❄➌❤➏ ➪✖➋❭➉❛✃❩➻ ➼Ñ➈☎➂➇➆★➚❭➈✪➏❛➂✚➈✆➑❦➋✥➒✇➒❸➋✥➪Ô➃❸➌✚Ð❞➆✥➎❁➎❤➃❸➏➍➃❸➋♠➌➇➆★➒➮Ð♠➈✍➌✚➈✍➉r➆★➒Ò➊✍➋❭➔✵➔✵➈✍➌❤➏❛➄✲➋♠➌ ➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋❭➌Ý➊❄➋♠➌❤➏❛➉❛➋❫➒✦➅❩➋✥➒✇➃❸➊➡➃❸➈✍➄✍➻✏➯➍➏➴➃❸➄✹➌✚➋♠➏✹➊r➒❸➈✥➆❫➉✂➂❁➋✥➪✽➅✚➉❛➈✍➊➡➃❸➄❛➈☎➄❛➐❁➊r➂ ➆❫➌⑧➆★➒❸Ð❭➋♠➉➍➃❸➏❛➂✚➔❐➌❁➈✍➈❄➎❁➄✹➏❛➋☎➶❩➈♠➻➮➯◗➑✦➏❛➂❁➈✍➉❛➈✹➃❸➄✹➈✍➌✚➋♠➐✚Ð♠➂▼➶➇➆❫➌✚➎✚➪Ô➃❸➎✚➏❛➂▼➏❛➋ ➔✵➈✍➈❄➏✯➔✵➋❭➄❛➏✯➊❄➐❁➄❛➏❛➋❭➔✵➈✍➉✂➌❁➈✍➈❄➎❁➄✍➹✦➆❫➌✚➎Ñ➃✇➑Ò➋♠➌❤➒❸➘→➆✆➄❛➔☎➆✍➒✇➒❳➑❦➉r➆❫➊✍➏➍➃❸➋❭➌▼➋❫➑ ➏❛➉r➆★Ó➾➊✆➌✚➈✍➈✍➎✚➄✂➏❛➂❁➈☎➔✵➋❭➄❛➏✯➎✚➈✍➔☎➆✥➌❁➎❤➃❸➌❁Ð⑧➋❫➑Ò➏❛➂❁➈☎➅✚➉❛➈✍➎❤➃❸➊✍➏❛➈✍➎❞➄❛➈✍➉❛➚♠➃❸➊✍➈♠➹ ➏❛➂✚➈✍➌❑➆❋➉❛➋♠➐✚Ð♠➂→➈✍➄❛➏➍➃❸➔☎➆❫➏❛➈❋➔➾➆★➘⑧➶❩➈✞➆❫➎✚➈✍➷❤➐➇➆❫➏❛➈♠➻→➯➍➌→➆❫➎✚➎❤➃❸➏➍➃❸➋♠➌☛➹✹➪✖➈ ➆❫➉❛➈✹➋✥Þ☛➈✍➉➍➃❸➌✚Ð❋➆✯Ð❭➈✍➌✚➈✍➉r➆★➒❱➔✵➈✍➏❛➂✚➋❤➎❋➪➴➂❤➃❸➊r➂✪➃❸➌❤➚♠➋❫➒❸➚❭➈✍➄✡✔✂➙➡➝♠➥❄❮❁➜r➢③↔❤➛❳➏❛➂❁➈ ➶❩➈✍➂➇➆✍➚♠➃❸➋♠➉✂➋❫➑✦➏❛➂❁➈✲➈✍➬♠➃❸➄❛➏➍➃❸➌❁Ð❋➉❛➈★➆★➒✇➓❏➏➍➃❸➔✵➈☎➏❛➉r➆★Ó☎➊❭➹☛➉r➆❫➏❛➂✚➈✍➉✹➏❛➂❩➆✥➌Ñ➐✚➄➍➃❸➌✚Ð ➏❛➂✚➈✹➏❛➉r➆★Ó☎➊✂➔✵➋❤➎✚➈➡➒❢➄❛➅❩➈❄➊➡➃✇➽❩➈❄➎ ➃❸➌❋➏❛➂✚➈✯➄❛➈❄➉❛➚♠➃❸➊✍➈P➃❸➌❤➏❛➈❄➉➍➑❏➆❫➊✍➈❭➹❩➃❸➌❋➎❁➈✍➊r➃❸➎❼➓ ➃❸➌✚ÐÝ➪❳➂✚➈✍➏❛➂✚➈✍➉☎➏❛➋▼➆✥➎❁➔✯➃❸➏✏➌❁➈❄➪✾➏❛➉r➆★Ó☎➊♠➻ ➼Ñ➈❋➐✚➄❛➈❋➏❛➂❁➈❋➪✦➋♠➉❛➄❛➏➍➓❏➊✥➆❫➄❛➈ ➏❛➉r➆★Ó➾➊✠➔✵➋❤➎✚➈➡➒❢➋❭➌❼➒❸➘☎➑❦➋♠➉✦➏❛➂❁➈Ò➌❁➈✍➪✫➄❛➋♠➐✚➉❛➊✍➈♠➹➇➪❳➂❼➃❸➊r➂✪➪✖➈Ò➊✥➆❫➌✚➌✚➋♠➏❳➋♠➏❛➂❤➓ ➈✍➉❛➪✖➃❸➄❛➈❋➊r➂❩➆✥➉r➆❫➊✍➏❛➈✍➉➍➃❸à❄➈✒✰✹➋♠➌✚➊✍➈❋➏❛➂✚➈❋➌❁➈❄➪ ➱➇➋✥➪✾➄❛➏r➆❫➉❛➏❛➄☎➉❛➐❁➌✚➌❤➃❸➌❁Ð✚➹✹➪✖➈ ➪✖➃✇➒✇➒❳➶❩➈✏➆❫➶❤➒❸➈☎➏❛➋✪➔✵➈★➆❫➄❛➐✚➉❛➈✪➏❛➂✚➈☎➆❫Ð❭Ð♠➉❛➈✍Ðä➆❫➏❛➈☎➏❛➉r➆★Ó☎➊☎➪✖➃❸➏❛➂▼➏❛➂✚➈✵➌✚➈✍➪ ➱➇➋❫➪→➆✥➌❁➎✵➶➇➆❫➄❛➈✖➑❦➐✚➉❛➏❛➂✚➈✍➉✿➆❫➎✚➔✯➃❸➄❛➄➍➃❸➋♠➌☎➎✚➈✍➊➡➃❸➄➍➃❸➋❭➌❁➄❳➋♠➌✂➏❛➂✚➈➴➔✲➋♠➄❛➏❢➉❛➈✍➊✍➈❄➌❤➏ ➔✵➈✥➆✥➄❛➐❁➉❛➈❄➔✵➈✍➌❤➏✍➻✹➁❳➂❤➃❸➄✂➆❫➅✚➅❁➉❛➋ä➆❫➊r➂Ñ➃❸➄➴➃❸➔✵➅➇➋♠➉❛➏r➆❫➌❤➏✹➑❦➋❭➉✠➏➍➪✦➋✆➉❛➈★➆❫➄❛➋♠➌✚➄✍➻ ✛✧➃❸➉❛➄❛➏✍➹❳➄➍➃❸➌✚➊✍➈Ñ➏❛➂✚➈❋➄❛➋♠➐✚➉❛➊✍➈✍➄✏➪✖➃✇➒✇➒✹➌✚➋♠➉❛➔☎➆✍➒✇➒❸➘❑➋♠➅❩➈✍➉r➆✥➏❛➈✪➃❸➌❁➄➍➃❸➎✚➈▼➏❛➂✚➈➡➃❸➉ ➒✇➃❸➔✯➃❸➏❛➄✍➹✂➏❛➂❤➃❸➄❋➪Ô➃✇➒✇➒✂Ð✥➃❸➚♠➈❞➆▼➶❩➈❄➏❛➏❛➈✍➉❋➊➡➂➇➆❫➉r➆✥➊✍➏❛➈✍➉➍➃❸à✥➆✥➏➍➃❸➋♠➌✽➆✥➌❁➎✫➶❩➈✍➏❛➏❛➈❄➉ ➒✇➃❸➌✚✃Ñ➐✚➏➍➃✇➒✇➃❸à✥➆❫➏➍➃❸➋❤➌☛➻✏➲❤➈✍➊✍➋♠➌✚➎☛➹✻➃❸➏✹➔☎➆❫➏❛➊r➂✚➈✍➄✹➪❳➂❩➆✥➏✠➏❛➂✚➈✵➊➡➒✇➃❸➈✍➌❤➏❛➄✂➏❛➂❁➈❄➔✯➓ ➄❛➈➡➒❸➚❭➈✍➄☎➆❫➉❛➈☎➎✚➋❫➃❸➌✚Ð❁➹✦➆❫➄✂➏❛➂✚➈✍➘▼➆✥➎❩➆✥➅❁➏✲➏❛➂❁➈☎➅❤➒❷➆★➘♠➓❏➶➇➆❫➊r✃▼➅❩➋✥➃❸➌❤➏✵➏❛➋❋➏❛➂❁➈ ➋♠➶✚➄❛➈✍➉❛➚❭➈✍➎Ñ➌✚➈✍➏❏➪✖➋♠➉❛✃✏➏❛➉r➆★Ó☎➊♠➻➮ß✠➆✍➚♠➃❸➌✚Ð✪➏❛➂✚➈✯➌✚➈✍➏❏➪✖➋♠➉❛✃ ➆❫➌✚➎Ñ➏❛➂✚➈✯➈❄➌❁➎❤➓ ➅❩➋✥➃❸➌❤➏❛➄❋➆❫➄❛➄❛➈❄➄❛➄✵➏❛➂❁➈❋➏❛➉r➆✍Ó☎➊✆➃❸➌→➄➍➃❸➔✯➃✇➒❷➆❫➉✏➪❳➆✍➘❤➄✯➃❸➄✵➒✇➃❸✃♠➈➡➒❸➘❑➏❛➋✞➶❩➈✍➏❛➏❛➈❄➉ ➅✚➉❛➋❤➎❁➐✚➊✍➈✂➉❛➈✥➆❫➄❛➋❭➌❩➆✥➶❼➒❸➈✵➶➇➈✍➂❩➆✍➚♠➃❸➋❭➉✍➻ Õ✒✑✔✓✠❽ ❚●✺❤❾ ✏❉✺❤❾✓✒♠Ü✇➀✔✺✖✕✹Û❤❿❃❂ Ü✇❽❛Ü✼✺♠❻ ➁❳➂❁➈❄➉❛➈✹➆❫➉❛➈Ò➆✹➌❤➐❁➔✂➶❩➈✍➉✦➋❫➑❢➋❭➏❛➂❁➈❄➉✖➄❛➈✍➉❛➚♠➃❸➊✍➈✠➑❦➈★➆❫➏❛➐✚➉❛➈✍➄✳➏❛➂➇➆❫➏✦➂❩➆✍➚♠➈Ò➶❩➈✍➈✍➌ ➅✚➉❛➋♠➅❩➋❭➄❛➈✍➎ ➃❸➌❋➏❛➂✚➈✹➊✍➋❭➌❤➏❛➈✍➬❤➏✠➋✥➑✿➉❛➈✥➆★➒✇➓❏➏➍➃❸➔✵➈✂➄❛➈✍➉❛➚♠➃❸➊✍➈✍➄✍➻Òß❳➈✍➉❛➈✹➪✖➈Ò➪Ô➃❸➄❛➂ ➏❛➋❋➔✵➈✍➌❤➏➍➃❸➋❭➌❞➏❛➂❁➈✍➔✏➹✿➆★➒❸➏❛➂✚➋♠➐✚Ð♠➂→➪✖➈✵➎✚➋Ñ➌✚➋♠➏✂➎❼➃❸➄❛➊❄➐❁➄❛➄✵➈❄➬✚➆❫➊✍➏➍➒❸➘▼➂✚➋✥➪ ➏❛➋✵➄❛➐✚➅❁➅➇➋♠➉❛➏✠➏❛➂✚➈✍➔✔➃❸➌❋➏❛➂✚➈✹➊✍➋♠➌❤➏❛➈❄➬❤➏✠➋❫➑❱➋♠➐✚➉✠➄❛➊r➂❁➈❄➔✵➈♠➻ ✫✠➌❁➈✵Ð❭➋❤➆✍➒✖➃❸➄✹➏❛➂❩➆✥➏P➃✇➑Ô➋❫➚❭➈✍➉➍➒❸➋❤➆✥➎❞➊✥➆❫➐✚➄❛➈✍➄✂➄❛➋♠➔✵➈✵➋✥➑❳➏❛➂✚➈✵➅❩➆✥➊➡✃❭➈✍➏❛➄ ➑❦➉❛➋❭➔❐➆✪➄❛➋♠➐✚➉❛➊✍➈☎➏❛➋✪➔✯➃❸➄❛➄✂➏❛➂✚➈➡➃❸➉✯➎✚➈✥➆✥➎❼➒✇➃❸➌✚➈♠➹✖➏❛➂✚➈☎➄❛➋♠➐✚➉❛➊✍➈☎➄❛➂❁➋❭➐❼➒❸➎❞➶❩➈ ➆❫➶❤➒❸➈✂➏❛➋✵➄❛➈✍➅➇➆❫➉r➆❫➏❛➈✹➃❸➏❛➄✠➅➇➆❫➊r✃♠➈❄➏❛➄➴➃❸➌❤➏❛➋☎➎❼➃✇Þ❙➈✍➉❛➈✍➌❤➏✹➊➡➒❷➆✥➄❛➄❛➈✍➄✍➹☛➏❛➋✵➊✍➋❭➌❤➏❛➉❛➋❫➒ ➪❳➂❼➃❸➊r➂▼➅➇➆❫➊r✃♠➈✍➏❛➄✹Ð❭➈✍➏✹➎✚➉❛➋♠➅✚➅❩➈✍➎☛➻✵➁❳➂❼➃❸➄✹➃❸➎✚➈✥➆✪➊✥➆❫➌✞➶❩➈✯➃❸➌✚➊✍➋❭➉❛➅❩➋♠➉r➆✥➏❛➈✍➎ ➃❸➌❤➏❛➋✵➋❭➐❁➉✖➄❛➊r➂✚➈✍➔✵➈✠➶❤➘☎➊✍➉❛➈★➆❫➏➍➃❸➌✚Ð✵➄❛➈✍➚♠➈❄➉r➆★➒✰➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘✏➊➡➒❷➆❫➄❛➄❛➈✍➄❳➪✖➃❸➏❛➂☎➏❛➂❁➈ ➄r➆❫➔✲➈✹➏r➆❫➉❛Ð♠➈❄➏ ✟✖ ➻ ➵❢➆❫➊r✃♠➈✍➏❛➄❳➏r➆❫Ð❭Ð♠➈✍➎✞➆✥➄✘✗➳➒❸➈❄➄❛➄➴➃❸➔✵➅❩➋❭➉❛➏r➆❫➌❤➏✚✙☎Ð♠➋✂➃❸➌❤➏❛➋ ➏❛➂✚➈✯➒❸➋✥➪✖➈✍➉✂➅❁➉➍➃❸➋❭➉➍➃❸➏➍➘▼➊➡➒❷➆✥➄❛➄✍➹✿➪➴➂✚➈✍➉❛➈✵➏❛➂✚➈✍➘✞➪✖➃✇➒✇➒Ò➆❫➉❛➉➍➃❸➚❭➈ ✲➳➐❁➄❛➏✹➶❩➈❄➂❼➃❸➌✚➎ ➏❛➂✚➈✵➔✵➋♠➉❛➈✂➃❸➔✵➅❩➋♠➉❛➏r➆✥➌❤➏✯➅➇➆❫➊r✃♠➈❄➏❛➄✍➹❢➶❁➐✚➏✹➪✖➃❸➏❛➂▼➂❤➃❸Ð♠➂✚➈✍➉✯➅✚➉➍➃❸➋♠➉➍➃❸➏❏➘▼➏❛➂➇➆❫➌ ➏❛➂✚➈✹➊➡➒❷➆❫➄❛➄❛➈✍➄✹➪✖➃❸➏❛➂☎➒❷➆❫➉❛Ð♠➈❄➉✞✟✖ ➻✿➯➍➏✳➃❸➄Ò➋♠➶❤➚♠➃❸➋❭➐❁➄✹➏❛➂➇➆❫➏❳➐✚➄❛➈✯➋✥➑❢➅✚➉➍➃❸➋♠➉➍➃❸➏➍➘ ➂✚➈✍➉❛➈✹➊✥➆❫➌❋➊❄➉❛➈✥➆❫➏❛➈✂➆✯➉r➆✥➌❁Ð❭➈✯➋✥➑❢➅❩➋✥➒✇➃❸➊➡➃❸➈✍➄✍➻ ☎✠➌✚➋♠➏❛➂✚➈✍➉✹➅❁➉❛➋❭➅❩➋♠➄❛➈❄➎Ý➄❛➈❄➉❛➚♠➃❸➊✍➈✵➃❸➄✹➏❛➂❩➆✥➏✹➅❩➆✥➊➡✃❭➈✍➏❛➄✯➏❛➂➇➆❫➏✂➆✥➉❛➈✵➄❛➐❤Ó✵➓ ➊➡➃❸➈✍➌❤➏➍➒❸➘✞➒❷➆❫➏❛➈✵➄❛➂✚➋♠➐❤➒❸➎❞➶❩➈✵➎❤➃❸➄❛➊✥➆❫➉❛➎✚➈✍➎▼➃❸➌❤➏❛➈✍➉❛➌➇➆★➒✇➒❸➘♠➹✿➉r➆❫➏❛➂✚➈✍➉✂➏❛➂❩➆✥➌Ñ➶➇➈➡➓ ➃❸➌✚Ð✆➎✚➈➡➒✇➃❸➚❭➈✍➉❛➈✍➎☛➹✻➄➍➃❸➌✚➊✍➈✠➃❸➌❋➎✚➈➡➒✇➃❸➚♠➈✍➉➍➃❸➌✚Ð❋➏❛➂✚➈✍➔✾➏❛➂❁➈✹➌✚➈✍➏❏➪✖➋♠➉❛✃☎➔☎➆✍➘☎➐❁➄❛➈ ➶➇➆❫➌✚➎❁➪✖➃❸➎✚➏❛➂Ý➏❛➂➇➆❫➏✹➊❄➋♠➐❤➒❸➎▼➂❩➆✍➚❭➈✵➶❩➈✍➈✍➌✞➶❩➈✍➏❛➏❛➈❄➉✹➐❁➄❛➈✍➎⑧➏❛➋✪➉❛➈✍➎✚➐❁➊❄➈✵➏❛➂❁➈ ➎✚➈➡➒❷➆✍➘✵➋❫➑☛➄❛➐✚➶✚➄❛➈✍➷❤➐❁➈❄➌❤➏✖➅❩➆✥➊➡✃❭➈✍➏❛➄✍➻❢➁❳➂✚➈➴➋✥Þ☛➄❛➈✍➏✿➊✥➆❫➉❛➉➍➃❸➈❄➎✯➃❸➌✵➏❛➂✚➈➴➅➇➆❫➊r✃♠➈❄➏ ➃❸➌❞➏❛➂❁➈✘✛✻➯■✛✽✫ ✰ ➄❛➊➡➂✚➈✍➔✵➈☎➅✚➉❛➋❫➚♠➃❸➎✚➈✍➄☎➅✚➉❛➈✍➊➡➃❸➄❛➈➡➒❸➘❞➏❛➂✚➈✆➌✚➈✍➈✍➎✚➈✍➎❞➃❸➌❼➑❦➋❭➉➍➓ ➔☎➆✥➏➍➃❸➋♠➌☛❒✳➃✇➑Ô➆✯➅❩➆✥➊➡✃❭➈✍➏Ò➆❫➊✍➊❄➐❁➔✂➐❼➒❷➆✥➏❛➈✍➄✂➆✯➚❭➈✍➉❛➘✆➒❷➆✥➉❛Ð♠➈✾✲➍➃❸➏❛➏❛➈✍➉❳➋❫Þ☛➄❛➈❄➏✍➹❩➃❸➏ ➃❸➄✦➆Ô➏r➆✥➉❛Ð♠➈✍➏❢➑❦➋❭➉✧➃❸➔✵➔✵➈❄➎❼➃❷➆✥➏❛➈✠➎❤➃❸➄❛➊✥➆❫➉❛➎❤➃❸➌❁Ð✚➻✖➁❳➂❤➃❸➄ ➃❸➎✚➈✥➆✠➂➇➆❫➄❢➶❩➈✍➈❄➌✵➅❁➉❛➋✥➓ ➅❩➋❭➄❛➈✍➎Ñ➈r➒❸➄❛➈✍➪❳➂❁➈✍➉❛➈❋➭✢✍ ✮✾✂ ✎❦➺❳➶❁➐✚➏✠➪✖➈✂➋♠➶✚➄❛➈✍➉❛➚❭➈✆➏❛➂➇➆❫➏❳➃❸➏✠➽❩➏❛➄✹➌➇➆❫➏❛➐✚➉r➆★➒✇➒❸➘ ➃❸➌❤➏❛➋☎➏❛➂❁➈✵✛✻➯■✛✽✫ ✰ ➄❛➊r➂❁➈✍➔✲➈♠➻ ☎✛➏❛➂❼➃❸➉❛➎❋➄❛➈❄➉❛➚♠➃❸➊✍➈P➃❸➄❳➏❛➂➇➆❫➏❳➅➇➆❫➊r✃♠➈✍➏❛➄Ò➄❛➂❁➋❭➐❼➒❸➎✞➶❩➈✹➶✚➐❼Þ☛➈❄➉❛➈✍➎✪➃❸➌❋➏❛➂❁➈ ➌✚➈✍➏➍➪✖➋❭➉❛✃➴➃✇➑➇➏❛➂❁➈❄➘✹➔✯➃❸Ð♠➂❤➏❢➋❭➏❛➂❁➈✍➉❛➪✖➃❸➄❛➈❳➆✥➉❛➉➍➃❸➚♠➈✖➈★➆❫➉➍➒❸➘✵➭❦➶❩➈➡➑s➋♠➉❛➈✿➏❛➂❁➈✿➅❼➒❷➆✍➘♠➓ ➶➇➆❫➊r✃✹➅❩➋❫➃❸➌❤➏r➺✻➄❛➋❳➏❛➂➇➆❫➏✻➏❛➂✚➈✿➈✍➌❁➎❤➓❏➌❁➋❤➎✚➈✖➌✚➈✍➈✍➎✂➌❁➋❭➏✻➅✚➉❛➋✥➚♠➃❸➎❁➈✖➏❛➂✚➈✿➶❁➐❤Þ☛➈✍➉➍➓ ➃❸➌✚Ð✹➋♠➉❢➈✍➄❛➏➍➃❸➔☎➆✥➏❛➈❳➏❛➂❁➈❳➊✍➐✚➉❛➉❛➈✍➌❤➏✿➎❁➈➡➒❷➆✍➘♠➻✿➼✞➈❳➆❫➉❛➈✦➌❁➋❭➏✳➊✍➋❭➌❤➚♠➃❸➌❁➊❄➈✍➎☎➏❛➂❩➆✥➏ ➏❛➂❤➃❸➄✖➄❛➈✍➉❛➚♠➃❸➊✍➈❳➃❸➄✳➐✚➄❛➈➡➑❦➐❤➒❩➃❸➌☎Ð♠➈✍➌✚➈✍➉r➆✍➒❦➻✳➼❞➃❸➏❛➂✪➊✍➐✚➉❛➉❛➈✍➌❤➏✿➔✵➈✍➔✵➋❭➉❛➘✯➊❄➋♠➄❛➏❛➄✍➹ ➶✚➐❼Þ☛➈❄➉➍➃❸➌❁Ð✵➎✚➋❤➈✍➄✖➌✚➋♠➏✿➄❛➈✍➈✍➔ ➈✍➬❤➅❩➈✍➌✚➄➍➃❸➚♠➈❭➻❏☎✠➌✚➎✵➪❳➂❤➃✇➒❸➈✠➃❸➏✳➔✯➃❸Ð❭➂❤➏✖➄❛➈✍➈✍➔ ➌❤➃❸➊✍➈ ➑s➋♠➉☎➏❛➂❁➈❋➌✚➈✍➏❏➪✖➋♠➉❛✃❞➏❛➋▼➉❛➈➡➒✇➃❸➈❄➚♠➈▼➏❛➂✚➈❋➎❁➈❄➄❛➏➍➃❸➌❩➆✥➏➍➃❸➋♠➌✫➈✍➷❤➐❼➃❸➅✚➔✵➈✍➌❤➏ ➑❦➉❛➋❭➔✔➏❛➂✚➈✠➌✚➈✍➈✍➎✪➏❛➋✂➈✍➄❛➏➍➃❸➔☎➆✥➏❛➈✹➏❛➂❁➈Ò➎❁➈r➒❷➆★➘❭➹❁➃❸➏❳➊✥➆❫➌✚➌✚➋♠➏✖➈➡➒✇➃❸➔✂➃❸➌❩➆✥➏❛➈✵➏❛➂❁➈
