276 Multiaccess Communication Chap.4 3.Collision or perfect receprion.Assume that if two or more nodes send a packet in a given time slot.then there is a collision and the receiver obtains no information about the contents or source of the transmitted packets.If just one node sends a packet in a given slot.the packet is correctly received. 4.0.1.e Immiediate feedhack.Assume that at the end of each slot.each node obtains feedback from the receiver specifying whether 0 packets.I packet.or more than one packet for error)were transmitted in that slot. 5.Retransmission of collisions.Assume that each packet involved in a collision must be retransmitted in some later slot.with further such retransmissions until the packet is successfully received.A node with a packet that must be retransmitted is said to be hacklogged. 6a.No buffering.If one packet at a node is currently waiting for transmission or colliding with another packet during transmission.new arrivals at that node are discarded and never transmitted.An alternative to this assumption is the following. 6b.Infinite set of nodes (m=x).The system has an infinite set of nodes and each newly arriving packet arrives at a new node. Discussion of assumptions.The slotted system assumption (1)has two ef- fects.The first is to turn the system into a discrete-time system.thus simplifying analysis. The second is to preclude,for the moment,the possibility of carrier sensing or early col- lision detection.Carrier sensing is treated in Section 4.4 and early collision detection is treated in Section 4.5.Both allow much more efficient use of the multiaccess channel, but can be understood more clearly as an extension of the present model.Synchronizing the transmitters for slotted arrival at the receiver is not entirely trivial.but may be ac- complished with relatively stable clocks.a small amount of feedback from the receiver, and some guard time between the end of a packet transmission and the beginning of the next slot. The assumption of Poisson arrivals (2)is unrealistic for the case of multipacket messages.We discuss this issue in Section 4.5 in terms of nodes making reservations for use of the channel. The assumption of collision or perfect reception(3)ignores the possibility of errors due to noise and also ignores the possibility of"capture"techniques.by which a receiver can sometimes capture one transmission in the presence of multiple transmissions. The assumption of immediate feedback (4)is quite unrealistic.particularly in the case of satellite channels.It is made to simplify the analysis.and we shall see later that delayed feedback complicates multiaccess algorithms but causes no fundamental problems.We also discuss the effects of more limited types of feedback later. The assumption that colliding packets must be retransmitted (5)is certainly reason- able in providing reliable communication.In light of this assumption.the no-buffering assumption (6a)appears rather peculiar.since new arrivals to backlogged nodes are thrown away with impunity.but packets once transmitted must be retransmitted until successful.In practice,one generally provides some buffering along with some form of flow control to ensure that not too many packets back up at a node.Our interest in this section.however.is in multiaccess channels with a large number of nodes.a relatively276 Multiaccess Communication Chap. 4 3. Collision or perfect reception. Assume that if two or more nodes send a packet in a given time slot, then there is a collision and the receiver obtains no information about the contents or source of the transmitted packets. If just one node sends a packet in a given slot. the packet is correctly received. 4. O./.e Immediate feedhack. Assume that at the end of each slot, each node obtains feedback from the receiver specifying whether 0 packets. I packet. or more than one packet (c for error) were transmitted in that slot. S. Retransmission of collisions. Assume that each packet involved in a collision must be retransmitted in some later slot. with further such retransmissions until the packet is successfully received. A node with a packet that must be retransmitted is said to be hacklogged. 6a. No huffering. If one packet at a node is currently waiting for transmission or colliding with another packet during transmission. new arrivals at that node are discarded and never transmitted. An alternative to this assumption is the following. 6b. Infinite set of nodes (m = x). The system has an infinite set of nodes and each newly arriving packet arrives at a new node. Discussion of assumptions. The slotted system assumption (I) has two ef￾fects. The first is to turn the system into a discrete-time system, thus simplifying analysis. The second is to preclude, for the moment, the possibility of carrier sensing or early col￾lision detection. Carrier sensing is treated in Section 4.4 and early collision detection is treated in Section 4.5. Both allow much more efficient use of the multiaccess channel, but can be understood more clearly as an extension of the present model. Synchronizing the transmitters for slotted arrival at the receiver is not entirely trivial, but may be ac￾complished with relatively stable clocks. a small amount of feedback from the receiver, and some guard time between the end of a packet transmission and the beginning of the next slot. The assumption of Poisson arrivals (2) is unrealistic for the case of multi packet messages. We discuss this issue in Section 4.5 in tenns of nodes making reservations for use of the channel. The assumption of collision or perfect reception (3) ignores the possibility of errors due to noise and also ignores the possibility of "capture" techniques. by which a receiver can sometimes capture one transmission in the presence of multiple transmissions. The assumption of immediate feedback (4) is quite unrealistic. particularly in the case of satellite channels. It is made to simplify the analysis. and we shall see later that delayed feedback complicates multiaccess algorithms but causes no fundamental problems. We also discuss the effects of more limited types of feedback later. The assumption that colliding packets must be retransmitted (5) is certainly reason￾able in providing reliable communication. In light of this assumption. the no-buffering assumption (6a) appears rather peculiar, since new arrivals to backlogged nodes are thrown away with impunity. but packets once transmitted must be retransmitted until successful. In practice, one generally provides some buffering along with some form of flow control to ensure that not too many packets back up at a node. Our interest in this section, however. is in multiaccess channels with a large number of nodes. a relatively