Final_exam_sample1_Solution - 1. (10) Multimedia over...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
1. (10) Multimedia over 802.11 Assume that your 802.11 MAC layer has been extended to provide two DFS inter-packet interval values, say DFS1 and DFS2, where DFS2 > DFS1. Suppose you are handling voice over IP, ie, you send voice packets over the 802.11 network. In addition to VoIP there is also TCP traffic. Questions: (a) 5 - which DFS value do you assign to voice, which to data (b) 5 - assume VoIP only traffic in your 802.11 network; moreover, assume that 802.11 has a capacity of 3Mbps (after the packet header and control overhead is excluded) and that each voice flow requires 30KBPS; can you safely state that the 802.11 network can carry 100 calls? Qualify your answer. Answer: a) You would assign the shorter DIFS1 to the VoIP traffic since it is more time sensitive. b) No, you cannot safely state that the 802.11 network can carry 100 calls. Since each flow requires 30Kbps (Note: Kilobits, not KiloBytes), the network would need to operate at 100% efficiency. This will not happen since terminals go into binary backoff after each transmission. Moreover, collisions between competing RTS can still occur, causing further backoffs and loss of channel efficiency. 2. (20) Link layer protocol with errors (a) 5 - why does the PPP data link layer protocol used in the Internet not provide ARQ (Automatic Request for Retransmission) and therefore error recovery? Which layer in the Internet does the detection/recovery of loss? (b) 5 - A possible exception to the above is a link with high error rate, say, a satellite link. For example, consider a satellite link with round trip time = 500 ms, bandwidth = 1Mbps, packet size = 5000 bits; packet loss due to errors = 1%. Suppose you are sending data and you decide to implement the Link Level sliding protocol with Selective Repeat; assume the LL sliding window size is selected so as to fill up the pipe. What is the window size? How many packets must the receiver be prepared to buffer? Why is Selective Repeat better than Go Back N? (c) 5 - If you now decide to do data loss recovery at the TCP level (instead of link level), how do you expect this to impact performance (assume TCP Reno with fast recovery)? A qualitative answer will suffice. (d) 5 - Now, in the satellite case, suppose you are transmitting not reliable data over TCP, but real time video over UDP. Do you think it is still a good idea to implement the Selective Repeat protocol for loss recovery?
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Answers: a) The PPP data link layer protocol does not provide ARQ and error recovery in order to increase efficiency and due to the low link error rates in the Internet. The transport layer will do detection/recovery of loss. b) In order to fill the pipe, we need a window size big enough so that we can transmit data for the entire RTT (500ms).
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 01/15/2012 for the course CS 111 taught by Professor Staff during the Fall '98 term at UCLA.

Page1 / 8

Final_exam_sample1_Solution - 1. (10) Multimedia over...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online