Question: Consider a network with end-to-end window flow control applied to each virtual circuit. Assume that the data link control operates perfectly and that packets are never thrown away inside the network; thus, packets always arrive at the destination in the order sent, and all packets eventually arrive.

(a) Suppose that the destination sends permits in packets returning to the source; if no return packet is available for some time-out period, a special permit packet is sent back to the source. These permits consist of the number modulo m of the next packet awaited by the destination. What is the restriction on the window size ~V in terms of the modulus m? Why?

(b) Suppose next that the permits contain the number modulo m of each of the packets in the order received since the last acknowledgment was sent. Does this change your answer to part (a)? Explain.

(e) Is it permissible for the source to change the window size ~V without prior agreement from the destination? Explain.

(d) How can the destination reduce the effective window size below the window size used by the source without prior agreement from the source? (By effective window size we mean the maximum number of packets for the source-destination pair that can be in the network at one time.)