《高级Web技术》参考资料：XMSF Technical Challenges Workshop Contributions.pdf_大学文库

1 Technical Challenges Workshop, Extensible Modeling and Simulation Framework (XMSF) Don Brutzman and Michael Zyda Naval Postgraduate School Mark Pullen, George Mason University (GMU) Katherine L. Morse, SAIC Science Applications International Corporation An Employee-Owned Company Monterey California USA, 19-20 August 2002 19-20 August 2002 2 XMSF Technical Workshop XMSF executive summary n Problems, motivation, strategy Workshop organization and details Administrative stuff Further technical discussion 19-20 August 2002 3 Executive summary Web-based technologies can provide an extensible modeling and simulation architecture, to support a new generation of interoperable applications Simulation support is needed for operational warfighting capabilities XML-based architecture can provide a bridge between emerging rehearsal/reality/replay requirements and open/commercial Web standards Web = best tech strategy + best business case 19-20 August 2002 4 Problems Current approaches are not compatible with effective use of emerging Web technologies Military modeling & simulation has little or no apparent impact on warfighters’ daily tactical operations Diverse simulations do not scalably interoperate with warfighting systems Global systems are not yet possible without connection to common interoperable framework ß physical and logical “stovepipes” prevent this 19-20 August 2002 5 Motivation Transformational technologies are needed to scale up defense modeling/simulation to meet real-world needs Can we use Web technologies as common framework? n Dynamic capabilities, open standards, Web business model provide lift to support government and commercial success n Easy use and open extensibility for developers and users, fueling rapid growth of interoperable simulations n Bring defense modeling/simulation/tactical support into mainstream of enterprise-wide best-business practices 19-20 August 2002 6 Planning for change/quick start Technical opportunities workshop n NPS, August 19-20 – welcome! n MOVES Open House, August 20-22 Strategic opportunities symposium nGeorge Mason University, September 6 n Immediately precedes SIW for good participation n Broader feedback: right track? what else is needed? Exemplar demonstration n I/ITSEC demos, Orlando Florida December 2-5

2 19-20 August 2002 7 Project strategy Outline large-scale framework for next-generation Web-compatible simulation interoperability Enumerate technical components, languages, and specifications n Workshop of key researchers to refine requirements n Seek broad feedback for consensus, focus, progress Demonstrate rapidly how these new capabilities might help n Current war effort against terrorism n Homeland defense n Science, commerce, education, etc. Multi-year initiative for simulation using Web technologies? n Workshop and symposium produce strategy and white paper n Conceptual suite of applications demo at I/ITSEC in December n NPS, GMU and SAIC have started, shared strategies welcome n Further workshop in fall among existing programs?? 19-20 August 2002 8 Workshop Details 19-20 August 2002 9 Workshop strategy 1 How to solve big problems? Divide & conquer. Three perspectives nWeb technologies, XML n Networking and Internet n Modeling and simulation (M&S) White paper lays out initial basic framework n please use, extend, correct, disagree 19-20 August 2002 10 Workshop strategy 2 How to solve big problems? Divide & conquer. Triage approach for all three technical areas: nWhat do we agree on: determine consensus nWhat do we disagree on: more work needed nWhat are most important directions for further work Document workshop and symposium results n Most important outcomes may be education, direction 19-20 August 2002 11 Web/XML group: Root 200C conference room n Moderator: Don Brutzman, NPS n Erik Chaum NUWC n Rob Glidden Sun n Jack Jackson, TRAC Monterey n David Kwak, MITRE n Recorders: Steve Fouskarinis SAIC, Curt Blais NPS n Dr. Francisco Loaiza, IDA n Dr. Edward Sims, Vcom3D n Dr. Chenghui Luo, Fraunhofer CRCG n Phil Zimmerman, DMSO 19-20 August 2002 12 Internet/networking group: Spanagel 254 conference room Moderator: n Mark Pullen, GMU n Rusty Baldwin, AFIT n Scott Bradner, IESG, Harvard n Suleyman Guleyupoglu, NRL n Sue Numrich, DMSO Recorders: n Don McGregor, NPS n Dave Laflam, AMSO n Denny Moen n Steve Carson, GSC Assoc. n Norbert Schiffner, CRCG n Marcelo Zuffo, University Sao Paolo n two last-minute drops: Sandeep Singhal Reefedge, Mikel Petty ODU

5 19-20 August 2002 25 Top-Level Requirements 1 Ability to interact directly and scalably over the network Compatible with Web architecture and technologies n Highly distributed Use by humans and software agents equally important Support for composable, reusable model components n Root data-structure representations specified using XML schema n Representations in other languages autogenerated directly n Connection point between syntax and RDF Schema, semantics 19-20 August 2002 26 Top-Level Requirements 2 Simple learning curve and repeatable examples n Support users and developers Modular structure n Ability to directly interact with network layer n Plug-ins connecting into kernel plug-ins at run time Standards-based n IEEE, ISO, W3C, IETF, Web3D Integrate with tactical systems n Augment group shared picture of operations n Producers and consumers n System life-cycle patterns, repeatability 19-20 August 2002 27 Top-Level Requirements 3 Support for XML and multiple programming languages Dynamically extensible at run time: “always on” n software + hardware, diversity includes backwards compatibility n loose coupling, verification/validation, repair, graceful degradation, redundancy, etc. Security levels consistent with current Web technology Public library of useful reusable components Cross-platform capabilities Rendering support and architectural hooks for visual simulations 19-20 August 2002 28 Top-Level Requirements 4 Expected computer performance: n Small, fast, inexpensive computers n Reconnect via GRID computing (distributed operating systems) Expected network performance: n Modems through ADSL (0.05-1.5 Mbps) for limited participation n 10 - 1Gbps for local participation n OC3 up through gigabit wide-area networking Backward compatibility with existing architectures and protocols, where it makes sense n e.g. DIS, HLA/RTI, ALSP, probably many others 19-20 August 2002 29 M&S Functional Requirements Backward compatibility Authoritative representations Composability Multi-resolution modeling Tactical system integration Simulation support services n Time management n Logging and playback 19-20 August 2002 30 M&S Issues Discuss the shared goal of bringing working M&S applications matching real world problems into tactical use. Discuss approaches for backwards compatibility to HLA/RTI and DIS technologies which don’t constrain emergence of new capabilities. Explore specific bridging approaches for HLA/RTI and DIS over web channels. Discuss compatibility with the Joint Technical Architecture (JTA), http://www-jta.itsi.disa.mil. Explore integration of C4I systems to augment joint common operational picture. Discuss approaches for playback capture. Identify technology availability: immediate, near-term (1-2 years), likely (3-5 years), problematic

6 19-20 August 2002 31 Design Ideas 1 Object-oriented programs + validatable structured data n Repeatable programming metaphors + XML tagsets/registries Language and object-system independent n e.g. CORBA/IDL, SOAP, RMI/RPC, others Design patterns unambiguously define language bindings n Map representations, component models from root XML schemas to programming languages and API bindings of interest Wire protocols unambiguously, flexibly defined using XML n Run-time extensibility, portability, interoperability for streams n Packet description language -or- direct mappability from Schema 19-20 August 2002 32 Design Ideas 2 Time services support wide-area routing and playback n Participating computer clocks all set correctly(!) via NTP or GPS n RTP for all stream headers n SMIL for stream synchronization n Higher-level time abstractions communicable by system n What about supercomputer batch-mode asynchronous? Infrastructure represented as objects in system n Enable object and service discovery at runtime n Ability to monitor, test, improve systems at local and global scale 19-20 August 2002 33 Business Model 1 Minimal architecture is open source implementation n Royalty free, usable without any fee restrictions whatsoever Important to have two or more interoperable implementations n Commercial implementations profitably augment open source Long-term stable infrastructure enables sustainable business models n Flexible architecture broadens market: not “just” military simulation, also full World Wide Web via open/secure Internets n Web-enabled architecture allows more sponsors to participate, which allows simulations, models, and applications to survive despite intermittent funding profiles n Transferable career-building skills and reusable experience for programmers and managers 19-20 August 2002 34 Business Model 2 Commercial models in partnership with open-source n Support w Offer programming skill for a fee and give up rights to the source if it is infrastructure related. n Proprietary software w A vendor may write a simulator that runs on top of the free infrastructure. They can use any license they like on this and sell it as they see fit. n Consulting w Someone needs to put together the simulations. Even if all the parts are free, this is a salable service. n Maintenance w Numerous success stories exist. Success of the overall endeavor is key to business success. 19-20 August 2002 35 DoD Business Model Commercial technology is crucial nWe can’t do it alone Special technology requirements harmful n Translate into always spending too much for unique, outdated technology Nothing succeeds like success n Slipstreaming standards and industry “best practices” makes best sense for industry partners too 19-20 August 2002 36 Open Issues Source code control n Interoperability test suite and certification shows conformance n Source forking no longer divisive issue, let many flowers bloom n “Actual results” validation through use with tactical systems Architecture n What are good driving exemplar applications? wReality pulls teams through the big design space wMust show significant “value add” to current practices plus current excitement n Can we extend some existing systems (e.g. HLA/RTI capable), show interoperability with legacy systems via good examples? we.g. NSS, OneSAF, ITEM n Multiple inheritance – no; multiple interfaces – yes n Do we need a specified method for distributed synchronization, or is open architecture supporting different models sufficient?