Tutorial #12

  1. What do you think is the origin of the term datagram?

  2. Why must a router always have at least two different IP addresses? Can a router have more than two IP addresses?

  3. A certain IP datagram encapsulates a TCP segment. The TCP segment, in turn, encapsulates portion of an HTTP GET request. Draw a diagram showing the boundaries between the various headers, and the location of the application data in the datagram. Assume minimum-sized headers.

  4. The three characteristics of datagram delivery in the Internet are:
    1. Unreliable
    2. Connectionless
    3. Best Effort
    Explain how each of these terms is interpreted in the context of datagram delivery.

  5. Explain briefly how a router reaches the decision to perform local delivery of a datagram. What is the other decision a router could make?

  6. Is[1] it possible to address a datagram to a router's (IP gateway's) IP address? Does it make sense to do so?

  7. A datagram travelling across the Internet is modified (ie, changed, altered) by every router that forwards it. How and why? Explain briefly.

  8. What is the function of a routing protocol in the Internet?

  9. The following is the output of a run of the traceroute command, "looking from" outside La Trobe back towards Bendigo. It was, in fact, run on a Unix system called morinda in the School of IT at NTU in Darwin.
    morinda> traceroute ironbark.bendigo.latrobe.edu.au
        [....seven lines deleted...]
     8  latrobe-gw.vrn.EDU.AU (  81 ms  83 ms  80 ms
     9  r-elt-fddi.latrobe.edu.au (  80 ms  81 ms  80 ms
    10  r-bgoatm34-atm.latrobe.edu.au (  81 ms  328 ms  84 ms
    12  busfddi0.bendigo.latrobe.edu.au (  81 ms  81 ms  81 ms
    13  ironbark.bendigo.latrobe.edu.au (  81 ms  81 ms  81 ms
    Use the information contained in this traceroute output to fill in the missing IP addresses in the "Internet Structure" diagram in the lecture.

  10. (Only for people who have already passed Discrete Maths or Computer Architecture 1) -- In the lecture it was stated that a host/router compares the network/subnet part of a packet's destination address with the network/subnet part of its own IP address. It's possible to do this with two Boolean operations -- first to extract the network part of each address, and then to compare them. What might these operations be?

[1] From Comer, Internetworking With TCP/IP, Vol 1, 3/e P.121.
La Trobe Uni Logo
Copyright 2004 by Philip Scott, La Trobe University.
Valid HTML 3.2!