Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!ucsd!swrinde!zaphod.mps.ohio-state.edu!samsung!dali.cs.montana.edu!milton!uw-beaver!cornell!ken
From: ken@gvax.cs.cornell.edu (Ken Birman)
Newsgroups: comp.sys.isis
Subject: Re: Bypass "configuration" for a trading floor problem
Message-ID: <44042@cornell.UUCP>
Date: 1 Aug 90 18:47:35 GMT
Sender: nobody@cornell.UUCP
Reply-To: factor-michael@CS.YALE.EDU (Michael Factor)
Distribution: comp
Organization: Cornell Univ. CS Dept, Ithaca NY
Lines: 53


   Article 359 of comp.sys.isis:
   Subject: Bypass "configuration" for a trading floor problem
   Message-ID: <43884@cornell.UUCP>

   I recently spent some time with a group that uses ISIS in a trading
   floor situation and was asked about the following.  The trading system
   this group is developing has a multi-tiered structure, in which
   various sorts of servers compute things (interest rates, market
   volatility, dividend projections, etc) and feed this information
   to one another in a sort of layered scheme.

	...

   I wonder if people have other architectures or problems we should be
   thinking about?  If so, please consider posting a brief description of
   your problem.  

I am a graduate student of David Gelernter's at Yale.  As my thesis
research, I have developed a software architecture, the process
trellis, for building parallel real-time monitors which seems similar
to the architecture you described.  This software architecture
organizes a collection of heterogeneous processes into a hierarchical
graph.  While the interpretation of the layers of the graph is not
fixed by the architecture a typical trellis will contain the following
layers: raw-data, data-filter, correlation, diagnosis, recommendation.
Data flows up the trellis hierarchy and queries (commands to calculate
new data) flow down the trellis hierarchy.

We have constructed with colleagues from the Yale Med. School a prototype
monitor for an intensive care unit that contains roughly 100 processes.
With colleagues from the NYU business school we have designed a trellis
program to monitor stock transactions and detect insider trading.

A process trellis shell provides a high-level interface for writing
trellis applications.  This shell implements all of the parallelism
and all of the inter-process communication, as well as scheduling the
processes on a multi-processor.  The goal of scheduling is to use the
fewest number of processors necessary to guarantee that the program
runs in real-time; our real-time guarantee is somewhere between hard
and soft real-time.  While this shell is implemented in Linda, it is
not tightly wedded to Linda, and there is no obvious reason it could
not be implemented in ISIS.

I have several papers on the process trellis which I would be happy
to send you if you are interested.

--- mike factor

INTERNET: <factor@cs.yale.edu>    	USMAIL: Dept Comp Sci, Yale University
BITNET:   <factor@yalecs.bitnet>		Box 2158, Yale Station	
USENET:   <factor@yale.uucp> 			New Haven, CT 06520
						203-432-1278