CS332 (Prof. Szajda) Exam 1 February 6, 2003

Name:______

Note: This exam is closed book, closed note. You are to complete without the aid of your colleagues. The Honor Code applies.

Please be succinct and precise.

  1. (6 points) What differences in traffic patterns account for the fact that STDM is a cost effective form of multiplexing for a voice telephone network and FDM is a cost effective form of multiplexing for television and radio networks, yet we reject both as not being cost-effective for a general purpose computer network?
  1. (5 points) Why is it important for protocols configured on top of Ethernet to have a length field in their header indicating how long the message is?
  1. (6 points) Explain why there must be a minimum frame length on an Ethernet. Be specific here: simply saying “minimum frame length is needed to detect collisions” (or the like) receives no credit. I want to know precisely WHY a minimum frame length is needed to detect collisions.
  1. (5 points) Suppose that N Ethernet stations, all trying to send at the same time, require N/2 slot times to sort out who transmits next. Assuming the average packet size is K time slots, express the available bandwidth as a fraction of N.
  1. (6 points) Why is it a bad idea to require a uniform maximum transmission unit for all networks running IP?
  1. (8 points) Write in C style pseudocode the (detailed) packet forwarding algorithm used by a router running the IP protocol. You need not include ARP issues in your pseudocode.
  1. (5 points) ARP is the protocol used to bind physical addresses with IP addresses. Let us assume that the underlying physical network is an Ethernet, and that a router on the Ethernet receives an IP packet that needs to be forwarded to another router on the same Ethernet. Give at least two reasons why it is not sufficient to simply broadcast the packet on the Ethernet. That is, give me at least two reasons why ARP is required at all?
  1. (15 points) Explain what each field in the first twenty bytes of the IPv4 header (see the figure below) represents, and how they are used. Be especially detailed when discussing those fields used for fragmentation. In particular, I want to know specifically how a destination node knows that it has received a fragment, how it knows that it has received a packet that has not been fragmented, and when it receives a packet that has been fragmented, how it knows when (and if) it has received all of the fragments.

  1. (10 points) List the advantages and disadvantages of distance-vector versus link-state routing. From a philosophical standpoint, what is the primary difference between the two?
  1. (6 points) What exactly is meant by saying that IP has a “best effort” service model? List at least three reasons why the best effort service model was adopted for IP.
  1. (5 points) IP hosts that are not designated routers are required to drop packets misaddressed to them, even if they would otherwise be able to forward them correctly. In the absence of this requirement, what would happen if a packet addressed to IP address A was inadvertently broadcast at the link layer? List two additional justifications for this requirement.
  1. (6 points) Give at least two examples of how protocol implementation issues significantly affect network performance. (The term implementation is in bold here to emphasize that I am not asking about protocol design issues here.)
  1. (6 points) Why is the management of memory so important for the performance of a network?
  1. (5 points) It is said that when acting as a network node, a workstation runs at memory, not processor speeds, to a first approximation. Explain.
  1. (6 points) In terms of system design, and especially in terms of network protocols, why is the idea of “centralized control” considered undesirable.