Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!rpi!batcomputer!cornell!ken
From: ken@gvax.cs.cornell.edu (Ken Birman)
Newsgroups: comp.sys.isis
Subject: Using ISIS as a teaching tool
Message-ID: <35067@cornell.UUCP>
Date: 11 Dec 89 20:19:31 GMT
Sender: nobody@cornell.UUCP
Reply-To: ken@cs.cornell.edu (Ken Birman)
Distribution: comp
Organization: Cornell Univ. CS Dept, Ithaca NY
Lines: 28

I've had a few questions concerning the use of ISIS in graduate
courses, and just wanted to post my thoughts on this.

First, I do this myself (and will be running such a course this
spring), and have had good success with it.  I usually run such a
course from papers and make copies of the manual available to the
students; an alternative would be to use a book like the one
on Distributed Computing from Addison Wesley's ACM Press Series,
which has two chapters on ISIS and covers a broad spectrum of topics.

I tend to give two types of assignments:

1) Type 1: implement a "classical" protocol using ISIS as an environment
   but not focusing on ISIS properties per se.
2) Type 2: implement an ISIS service of some sort.

Once ISIS V2.0 is out next spring, type 1 assignments should give
performance pretty close to the "bare wire" numbers, which is 
attractive for performance evaluation too.  Needless to say, the ISIS V1.3
picture is a bit different.

It strikes me that all of us who are doing this sort of thing might
benefit from sharing our experiences... I propose that we begin
to post homework problems and how they went over and that we run
a subtopic discussion on this.  Everyone should benefit, and we'll
save each other the trouble of making up a huge number of assignments.

Ken