CS231 Lab 4

Out date : 11/11/02
In date : 11/29/02

Distributed Clock Synchronization

Object : Distributed Clock Synchronization Using Vector Clocks and
Casual Message Ordering by Birman et all Protocol

Procedure : You will implement first the Vector Clocks and then the
Birman Protocol . The nodes will be simulated by threads communicating
via FIFOs and managing buffer space by buffer pools.


A. Vector Clocks

A.1. Construct the implementation rules for Vector Clocks (Section 5.4.
of the text book)
A.2. Provide 3 Solaris kernel threads T1,T2,T3 such that :

T1 : Performs four events and sends a message to T3
T2 : Sends a message to T1, performs two events then sends a
message to T3
T3 : Sends a message to T2, sends a message to T1, performs an
event, then sends a message to T2

A.3. Show the operation of the threads and log the vector clocks in a
file for processing.
A.4. Repeat for other random message passing

B. Casual Ordering Of Messages

B.1. Construct Birman-Schiper-Stephenson Protocol for each thread
B.2. For the A.2. example, provide a broadcast FIFO for all or use
each thread FIFO to implement Birman et all protocol.
B.3. Show that the protocol works for random messages order of which is
entered from the keyboard.

Grading : Correctness 20 %

Completeness 20 %

Documentation 10 %