From shasha@SHASHA.CS.NYU.EDU Wed Jun  5 13:40:52 1991
Return-Path: <shasha@SHASHA.CS.NYU.EDU>
Received: from corton.inria.fr by madonna.inria.fr. (4.0/SMI-4.0)
	id AA18035; Wed, 5 Jun 91 13:40:50 +0200
Received: from SHASHA.CS.NYU.EDU by corton.inria.fr (5.65c+/90.0.9)
	via Fnet-EUnet id AA26874; Wed, 5 Jun 91 13:57:29 +0200 (MET)
Received: by SHASHA.CS.NYU.EDU (5.61/1.34)
	id AA02442; Wed, 5 Jun 91 07:56:52 -0400
Date: Wed, 5 Jun 91 07:56:52 -0400
From: shasha@SHASHA.CS.NYU.EDU (Dennis Shasha)
Message-Id: <9106051156.AA02442@SHASHA.CS.NYU.EDU>
To: shasha@madonna.inria.fr
Subject: justification
Status: R

General Justification of Dennis Shasha's existence

Unlike most of our contractors, Shasha has another professional
life that contributes to his work at USL.
He is a tenured professor with an active and well-published
research effort in the theory and practice of distributed systems:
transaction processing systems, fault tolerant parallel systems,
and distributed systems.
This means that his other work (which we don't pay for) contributes
to his effectiveness here.

For example, his work on real-time scheduling theory will help
inform our design of priority facilities in the kernel.
To give another example, his work on transactional Linda will
help his understanding of the support we need for transactions
and parallelism.

Finally, his contacts in academia can help bring needed
ideas into USL.

His concrete contributions: his memo on micro-kernels
helped set the stage for much of the work the FLW group
has done this year.
He has been an active and, by all accounts, co-equal partner
in the FLW design effort this year.
His understanding of distributed algorithms will make him
equally useful next year as the implementation project
gears up.

Money we spend for him: he also consults on occasion
for companies on Wall Street and has served as an expert
in several legal cases.
In those situations, he receives 150 dollars per hour,
nearly twice what we pay him. 
He works for us in spite of this difference because he views
his work here as supportive of his research and teaching.
The differential is large as it is, however, so our lowering
his wage may drive him away. 


Specific activities in France:

There are some advantages to Shasha's being in France
and some disadvantages.

Advantages:

1. He will be near to large numbers of people who use Chorus:
a. researchers at Inria (where he will be working and
where Chorus was first developed) use Chorus in several projects
b. he may do some work with some real-time groups working
for the European Space Agency, another user of Chorus.

2. He can go to Chorus to run benchmarks.
This was difficult for Jishnu without active support.
Also, this will help establish personal contacts 
in the event of future collaboration. 

3. He will continue his research in distributed fault-tolerant
systems, thus informing his recommendations to us.

Disadvantage:

He will not be able to participate in face-to-face meetings,
though he will be able to participate in conference calls
and email discussions.

Activities in France:

1. Run Gaede on Chorus.

2. Look at various implementations that use Chorus.

3. Study various models of distribution and fault tolerance
to see how they are reflected in the kernel, particularly
Linda, ISIS, and transactions (on databases, on file systems,
and on virtual memory).

4. Review and participate in the ongoing design and implementation of the
kernel model and its execution model.

5. Review the distributed computing architecture.

.bp


