Question 1 The UDP header does not contain any information that is not present in the TCP header (verify this for yourself). That being the case, why do you, as a programmer, have to specify the header information (ports and IP addresses) whenever you send a UDP packet; but you don't when you send a TCP packet?

Question 2 Protocol design decisions often have unexpected perfomance consequences. HTTP 1.0 is an example of a protocol design where lower layer protocol behaviour impacted directly on the performance of the higher layer protocol. What is the difference between HTTP 1.1 and HTTP 1.0 in terms of transport layer connections? What two transport layer issues do the changes to HTTP address? How does the change improve HTTP performance?

Question 3 Consider what you have learned about congestion control in TCP Reno (most common algorithm). How might application designers exploit the Internet's use of TCP to get higher data rates at the expense of other data flows that are using TCP?

Question 4 Consider the problem of implementing timers in your current practical. Selective Repeat does not resend all packets on timeout, so it must timeout packets individually. However, in our implementation we only have one timer. How might we solve this problem? What would happen if we re-start the timer every time a packet is sent? What would happen if we re-start the timer when the oldest packet is ACKed?