Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!lll-winken!uunet!ssbell!mcmi!denny
From: ac@k.gp.cs.cmu.edu (Alan Chung)
Newsgroups: news.announce.conferences
Subject: CFP: Hawaii International Conference on System Sciences
Message-ID: <918@mcmi.UUCP>
Date: 7 Apr 89 03:00:22 GMT
Expires: 14 Apr 89 23:00:00 GMT
Sender: denny@mcmi.UUCP
Reply-To: ac@k.gp.cs.cmu.edu (Alan Chung)
Organization: Carnegie-Mellon University, CS/RI
Lines: 89
Approved: denny@mcmi


[See also <848@mcmi.UUCP> of 03/17/89  -mod]


			   CALL FOR PAPERS
	  HAWAII INTERNATIONAL CONFERENCE ON SYSTEM SCIENCES
			       HICSS23

		Kailua-Kona, HAWAII - 2-5 January 1990

The architecture track will contain a minitrack on

		 PERFORMANCE EVALUATION AND MODELING

The role of performance models is to predict potential bottlenecks
during the design process, and thus reduce development risks and
design effort. Once a system is built and under test, models can also
facilitate characterization of its behavior. It would be ideal if
models could predict and characterize all potential bottlenecks, but
is is usually satisfactory in practice to obtain only realistic bounds
on performance.

The issue of performance evaluation is extremely critical in
distributed and parallel systems. Modern parallel systems are designed
to achieve two main goals: high performance and increased
availability. Both goals can be achieved via parallel use of system
resources, but one should be aware that the use of parallelism
increases system complexity. While abstractions simplify design-time
complexity, they can be a major source of run-time performance
bottlenecks. These bottlenecks usually appear in one of four forms:
delays due to contention on common resources, delays due to
synchronization overhead, increased load due to unfavorable parallel
decomposition, and unbalanced load on the resources in the system.

Papers are sought that will cover theoretical and implementational
issues related to performance efficient design of parallel,
distributed, fault tolerant and real-time systems. They may be
theoretical, conceptual, tutorial or descriptive in nature. Those
papers selected for presentation will appear in the conference
Proceedings, which is published by the IEEE Computer Society. HICSS-23
is sponsored by the University of Hawaii in cooperation with the ACM,
the IEEE Computer Society, and the Pacific Research Institute for
Information Systems and Management [PRIISM].

Papers are expected to cover various fields of performance evaluation
including:

    - New theoretical performance models.
    - Performance of parallel/distributed systems.
    - Novel applications in high performance computing.
    - Automated tools for performance efficient design.
    - Performance monitoring and diagnostics.
    - Queuing networks and their solutions.
    - Applications of Petri nets.
    - Performance of Real-Time systems.
    - Approximate solutions.
    - Deterministic models.

Instructions for Authors:

Submission should be no longer than 22-26 double-spaced pages
including all text, figures and appendices. They must be original
works and will be submitted to a reviewing process. Each submission
should have a title page which contains a title and names of authors
with their addresses and telephone numbers.

There are the following deadlines for submission process:

	A 300-word abstract:		April 15, 1989.

	Feedback to the author:		May 1, 1989.

	Six copies of the manuscript:	June 5, 1989.

	Notification of acceptance:	August 31, 1989.

	Camera-ready copies of
	accepted papers:		October 1, 1989.


PLEASE SEND YOUR SUBMISSIONS OR DIRECT YOUR QUESTIONS TO:

		       Dr. Dalibor F. Vrsalovic
		      School of Computer Science
		      Carnegie Mellon University
		      5000 Forbes Ave. /WeH4104
			Pittsburgh, PA 15208.
			 tel: (412) 268-6426
		       e-mail: dv@@k.cs.cmu.edu