Assignments for MSCS209, Fall 1998

Final Exam

due on Friday, December 11, 1998, by 11:59 PM.

I decided to give you another decade!
Especially since no one turned it in in 1988.
The exam is a list of terms, for which you should provide a paragraph or two of explanation, in your own words, as gleaned from lectures and the Web.
  • You should edit the file final.html, which is just a definition list.
  • It is enclosed in <DL> tags:
  • I have provided the <DT> entries,
  • and an empty <DD> tag with each.
  • Thus I have done half the work already:-).
  • Your task, which you must choose to accept,
  • is to fill in the DDs with illuminating words.
You are welcome to provider your own HREFs to additional material, but you should provide a synposis yourself.

Grading

Each of the 6 assignments will be 10 points, and the final will be 40 points.

However if there is a particular assignment that really stands out, I will substitute your grade on that assignment for the one that is the worst (in my opinion).

This is similar to "I will drop your lowest grade", but adds "I will double your highest grade"!

Base

Make sure you have a home page:
this will be a file


/home/LOGNAME/public_html/index.html
accessible by the URL

http://studsys.mscs.mu.edu/~LOGNAME


Make sure you have a class directory,


/home/LOGNAME/public_html/MSCS209/
which will contain an index page,

index.html
, and your assignment pages
  1. 
    Assignment01.html
    
    ,
  2. 
    Assignment02.html
    
    ,
. and whatever material you wish to place in it.

Either your home page or your class page should contain a scanned photograph of yourself (no substitutions!).


Assignment 01, due September 23, 1998

Part I

Use snoop to capture an ARP transaction, both request and reply.
For example, on netlab-a:

snoop -v arp > myarpcapture &
arp -d spectral
ping spectral
kill %1

If you do not tell it to capture only arp, then edit your file
to contain only the ARP request and reply.

Part II

Use snoop to capture a RARP transaction, both request and reply.
Since you cannot easily generate a RARP request yourself, an X terminal is installed in the networks lab, and when you reboot it (turn it off and on again) it will do a RARP request, which you save for the assignment.
For example, on netlab-a:

snoop -v > myrarpcapture &
[ ... Now turn the X-terminal off and on again ... ]
kill %1

In this case you will have to edit the file,
since it will contain things other than RARP.

Assignment 02, due October 7, 1998

  1. On the mailing list comp.dcom.cell-relay, someone said they had heard that "with ATM it is important that the cell loss probability be very low, because a very small number here could result in large packet loss". Some members of the list explained why, others explained why not. Pretend I am your boss, I heard this statement, it had me worried, and I wanted you to settle it for me in a way I could understand. If you want to try some sophisticated statistics, fine, but make sure the energy you spend is worth it. Simplicity will be appreciated: note however the question is not trivial.
  2. Explain in your own words the differences between
    • Classical IP over ATM
    • LAN Emulation over ATM
  3. Explain in your own words the differences between
    • Reverse ARP
    • Inverse ARP
  4. In the ATM over IP framework RFC1932, toward the end, there is a description of how a routing loop might be formed. I did not understand the description. Read it yourself, and see if you can explain it, and construct a simple example to illustrate it. Graphics would be appreciated if you can do them, but don't kill yourself, or me.

Assignment 03, due October 21, 1998

I am still your boss, and now I have "moved up" the ISO/OSI layers a little, and am hearing things that I want you to explain to me at that level.

I heard about a battle that raged for a while, involving the Internet providers BBN and Exodus (both are huge).

BBN in this scenario seemed to be acting as an "access provider", and Exodus as a "content provider", or at least a "content host".

That is, Exodus hosts some very popular Web sites. Web traffic is characterized by very short queries from the client, very long replies from the server. It is also characterized by many individual connections of client to server, one in fact for each "file" that needs to be accessed: thus every image on a page, even a little button, involves a connection.

BBN seems to have noticed that they were giving Exodus only a small amount of data, and Exodus appeared to be dumping everything to them. This apparently made someone think that the "peering" was not fair, and they took steps to do something about it, namely to try to charge Exodus for the traffic.

Exodus responded at least in part that they were being "good citizens", and attempting to use their own net to deliver the traffic as long as they could. In particular they were using "last exit" rather than "first exit" to transfer their traffic to BBN.

What do you think? Can you set out some issues so that I can understand this?

Assignment 04, due October 28, 1998

You will be asked to write an RIP provider in Java. Here is code for most of the listener part of the project.

Take a look at RIPsimulate for the helper programs you need.

Assignment 05, due November 30, 1998

Implement the TFTP protocol in Java.
There should be three applications (i.e., that can be run as a "main"):

Assignment 06, due November 30, 1998

Use "traceroute" to study internet paths.

Use altavista or another search engine to find Web hosts that will run a "return traceroute" for you.

Run a traceroute to a variety of such hosts, and have them run a traceroute to you at more or less the same time.

See if you can explain any variations in paths taken, or times required, and so on, based on time of day, location, or other relevant factors.

Since this will be done at different times of day, it is OK if several of you work together. Be sure to list all the members of your group, and realize that I expect 4 times as much from a group of 4 (and even a little more for synergy!).


Prepared by     doug@mscs.mu.edu Douglas Harris
Created September 09, 1998