Xref: utzoo ont.events:1417 uw.talks:111 uw.cs.grad:89
Path: utzoo!attcan!utgpu!watserv1!watmath!maytag!water!wlrush
From: wlrush@water.waterloo.edu (Wenchantress Wench Wendall)
Newsgroups: ont.events,uw.talks,uw.cs.grad
Subject: MASTER'S ESSAY PRESENTATION
Keywords: Mr. Debasish Jana, graduate student, Dept. of Comp. Sci.
Message-ID: <2861@water.waterloo.edu>
Date: 18 Dec 89 19:54:31 GMT
Distribution: ont
Organization: U of Waterloo, Ontario
Lines: 59


DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES

MASTER'S ESSAY PRESENTATION

                    -Thursday, December 21, 1989

Mr.  Debasish Jana, graduate student, Dept. of Computer
Science  will  speak on ``Synchronous Communication and
Static  Process  Structure  -  Problems  and a Proposed
Solution.''

TIME:                 2:30 p.m.

ROOM:                 DC 1331

ABSTRACT

A   programming   language   for   distributed   system
environment  has  to  choose  a  set  of  communication
primitives(synchronous    or   asynchronous)    and   a
structure of processes(static or dynamic).

In a client-server model, a client requests for service
from a server and the server, in turn, requests several
resources for their availabilities. A server encounters
a  local  delay  when  a  local  resource needed is not
available  and  it  encounters  a remote delay when the
server  makes a call to another module which is busy or
has  to  perform considerable computing before making a
response.   Remote   delays   are   mainly  because  of
communication  or  I/O delays. Liskov et al argued that
the   combination   of  static  process  structure  and
synchronous communication has adequate expressive power
for  handling  local  delays.  But,  it is not entirely
suitable  for  dealing  with  remote  delays. They also
suggested   that  a  distributed  programming  language
should  abandon  either  synchronous  communication  or
static process structure.

In   this   essay,   we   propose  a  new  facility,  a
communication primitive reply-with that shows Liskov et
al's   hypothesis   is  not  entirely  true.  This  new
primitive    largely   removes   the   limitations   of
synchronous  communication  primitives  in  combination
with static process structure, when combined with send,
receive, reply (Thoth's message passing primitives) and
also   with   other  facilities  e.g.  Ada  rendezvous,

                   December 13, 1989


                         - 2 -

monitors.

                   December 13, 1989