CPS 214 Written Homework 4 - Fall 1997

Due: Wednesday, November 12, 5:30 PM

10 points

  1. In the ARP algorithm described in class, a booting node sends out an ARP packet asking for its own address. If the booting node is attempting to use an IP address that is a duplicate (i.e., another node is already using that address), the other node will respond to the ARP request and the booting node can then decide not to use that address. Is it too late at that point? That is, if the booting node decides not to use the duplicate address, have neighboring ARP caches already been corrupted with bad data?
  2. Explain how sending IP packets to a nonexistant destination address on a remote Ethernet can generate excess broadcast traffic on that network.
  3. In IP Version 6, the checksum has been removed from the IP packet header. What happens if one of the bits in the header of an IPv6 datagram becomes corrupted? Is the result catastrophic?
  4. Before sending an ICMP error in response to an error encountered while attempting to forward a unicast IP datagram, nodes are required to check that the datagram causing the problem was sent to a unicast rather than multicast (or broadcast) address at the link-layer. In the latter case, no ICMP error is sent. Why do you think such a rule was specified?
  5. NCREN has been assigned 60 Class B addresses (152.1/16 - 152.60/16) to be used by the sites that connect to it. Through aggregation, is it possible for all NCREN sites to be reachable through a single CIDR route prefix, or should they have been assigned a different range of addresses? Explain.
  6. Explain why it is necessary to have one address per interface, rather than just one address per host.
  7. The Version field is the first field of the IP header. Is this placement arbitrary? That is, could the version field appear elsewhere, such as at the end of the header?
  8. Suppose that a gateway was fragmenting an IP datagram (carrying a TCP segment) but neglects to set the ``fragment more'' bit in any of the fragments. What is the likely consequence? Will the remote TCP module process the packet and accept the wrong data or will something else take place? Explain.
  9. Is ARP used when sending IP multicast traffic?