6/73201 Advanced Operating Systems
			PROJECT 1 ASSIGNMENT
			   50 points

        Due by e-mail at 11:59pm on Tuesday, September 25, 2007

			 RANDOM FLOOD

Assume that the system is completely connected. The execution
semantics is interleaving and the system is message passing.
Processes have unique identifiers ranging from 0 to N-1 (where N is
the number of processes in the system).

You are to implement an algorithm that floods a message to all
processes in the system.  The algorithm operates as follows. A process
with identifier 0 initiates the algorithm by executing a random
broadcast. This broadcast entails selecting a random number of
neighbors and sending them a message. The message bears no content.
The number of random neighbors is uniformly distributed from 0 to N-1.
Note that the number of selected neighbors may thus be 0.  The
recipient identifiers are randomly selected from the list of all
identifiers. When a process receives a message for the first time, it
executes a random broadcast of its own. When this process receives
consequent messages, no further actions are executed. The algorithm
terminates when there are no more messages to receive. Note that
the initiator, like all other processes, executes the random
broadcast at most once.


		      SIMULATION ENGINE

The important part of the project is the simulation engine: the
portion of code that simulates message transmission. Make sure this
portion is modular and can be reused in further projects. The
simulation engine should keep the state of each process and the queue
of messages in transition (currently in the channels). The engine
would construct a computation of the algorithm by iterating through
the messages in the queue. The engine would select the message at the
head of the queue, and have the recipient process execute a single
atomic step. This execution would result in updating the state of the
process and possibly some messages to be sent. The messages shall be
appended to the tail of the queue.


		    IMPLEMENTATION DETAILS

For you implementation you can use either Java or C++. Your code
should compile and run on the departmental Unix servers.  If you
prefer to use any other language, please discuss it with me before
starting working on the project. Make sure you properly initialize the
seed for your randomization implementation. I suggest initializing it
from the computer's clock. Make sure your implementation is
configurable to different values of N (the number of processes in the
system). This configuration can be done through a run-time parameter
or a compile-time constant. Coding has to be done professionally: with
comments, indentation and appropriate modularity. Your grade may
be influenced by the quality of your code. 


		    SUBMISSION DETAILS

Pack and send your project to Tom Clouser (see class web-age for
email). Your submission should contain the source code and
instructions for compiling and running it. If you project does not
work, send a note explaining what stage it is at and why you think it
does not work.  Tom will acknowledge your submission after he is able
to run your code. Late submission policies apply as in the
syllabus. Note that grading will be such that submitting a working
project a few days late will get a higher grades than submitting a
non-working project on time.