Xref: utzoo ont.events:1456 uw.talks:135 uw.cs.grad:121
Path: utzoo!utgpu!watserv1!watmath!maytag!water!wlrush
From: wlrush@water.waterloo.edu (Wenchantress Wench Wendall)
Newsgroups: ont.events,uw.talks,uw.cs.grad
Subject: DISTRIBUTED SYSTEMS SEMINAR
Keywords: Prof. S. Singh, Dept. Computer and Information Science,
Message-ID: <2949@water.waterloo.edu>
Date: 1 Feb 90 13:56:59 GMT
Distribution: ont
Organization: U of Waterloo, Ontario
Lines: 48

University of Massachusetts, will speak on 
`` Preference-Based Leader Election in Distributed Systems.''

DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES

DISTRIBUTED SYSTEMS SEMINAR

                    -Thursday, February 8, 1990

Professor   Suresh   Singh,   Dept.   of  Computer  and
Information  Science, University of Massachusetts, will
speak   on   ``Preference-Based   Leader   Election  in
Distributed Systems.''

TIME:                 3:30 p.m.

ROOM:                 DC 1304

ABSTRACT

In  a distributed system, an algorithm used to select a
distinguished   node   or  leader  to  coordinate  some
                           ------
activity  in  the  system is known as a leader election
                                        ------ --------
algorithm.   Existing leader election algorithms may be
----------
characterized  as  extrema-finding  algorithms in which
all  nodes  are assumed to have a unique ID number, and
the  leader  which is elected is simply that node which
has  the  largest ID number.  In this paper, we examine
socially-inspired   leader   election  algorithms  that
attempt  to  locate  the  centralized  controller  at a
``good''  node  (from  a performance standpoint) in the
system.    In   the   ``preference-based''   approaches
examined  here,  each  node  in the system uses locally
available  information  to  ``vote''  for  the  various
candidates  (potential  leaders)  on  the  basis of the
performance  level it would realize under each of them.
The  preference-based  leader  election  algorithms  we
propose  and  examine  are  simple, are shown (via both
simulation  and  analysis) to perform almost as well as
traditional  optimization-based approach towards leader
election,  and  have  important  advantages  over  both
current    leader    election    algorithms    and   an
optimization-based approach towards leader election.