Suppose that the TCP connection between A and B goes through router R (A to R then R to B). A-R link has infinite bandwidth and zero delay; R-B link has infinite bandwidth but 1sec propagation delay. Router R has infinite buffer size while node B has a finite buffer size of 8 packets. Node B's application reads one packet out the buffer every 2sec, but it only reads data in sequence. For example, if P1, P3,P4 are in its buffer, B's application will not read P3 after reading P1 because P2 is not present. Node B sets the advertised window as at most half of its available buffer size. For example, if there are 3 unread packets at B's buffer, the advertised window equals to floor{ (8-3)/2} = 2. The information of advertised window size is piggybacked on ACK. Router R instantly forwards packets and ACKS to B and A, respectively. B only ACKS the packet of highest sequence number. For example, if P1, P2, P3 are received by B, B only returns ACK3. If P2 is lost while P1,P3 arrive correctly, B sends ACK1 back. Host A processes information in the following order: first process any ACKS, then process any timeouts and then send more packets if possible. The whole processing takes Osec. Host A always sets its sending window size equal to the advertised window size, even when timeout occurs. The initial sending window is 1 packet. The timeout time is 4 seconds. Suppose B's application starts to read out data at t=1sec. If the first packet (P1) sent by Host A occurs at time t=1sec, and P3 is lost in its first transmission attempt, please answer following questions. a) (3pts) At what time does a timeout of a packet first occur? b) (3pts) What is the sending window size at time t = 5sec? c) (3pts) At what time does the sending window size drop to 1 packet? d) (3pts) Is there any chance that B's buffer will overflow? If yes, at what time? If not, explain why.