Consider T-persistent CSMA. The channel is time-slotted with a time slot lasting one propagation delay (tau), as shown in the figure below. Assume the same assumptions we made for the discussion of slotted ALOHA and slotted CSMA. Namely: Poisson arrivals with parameter lambda, fixed packet lengths of duration P seconds, a fully-connected network with all nodes having tau seconds of propagation delay from each other, and a magical secondary channel over which ACKs are sent without error in zero time. In contrast to non-persistent CSMA, a node that has a packet to send and detects carrier does not back off immediately as in non-persistent CSMA. Instead, it "persists for T seconds." This means that, after detecting carrier, a node with a packet to send continues sensing the channel for T seconds [a multiple of the propagation delay tau] and transmits its packet at the next time slot that the channel is free within the T seconds. The following figure illustrates the operation of the protocol when T = P +tau Note that there are three states (type of transmission periods) defined by the number of arrivals in the last T seconds of the prior transmission period, again with T = P + tau in the example shown in the figure. (a) Using a similar approach to the one we discussed for slotted ALOHA and slotted CSMA compute the throughput of the protocol when T = P + tau. Be careful! Notice the difference in transition probabilities from idle to busy (TP1 or TP2)or idle, and from busy (TP1 or TP2) to idle (TPO) or busy CTP1 or TP2). (b) Discuss not compute] what changes you would need to do to the analysis in order to consider the existence of acknowledgments being sent over the same channel. (c) Repeat your analysis for the case in which T = tau. Does the result look familiar? What pros and cons can you state between T = P + tau and T = tau