Xref: utzoo bionet.general:1160 sci.research:1489
Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!ohstpy!miavx1!miamiu!js05staf
From: JS05STAF@MIAMIU.BITNET (Joe Simpson)
Newsgroups: bionet.general,sci.research
Subject: Re: Directions for scientific computing centers for the future
Message-ID: <90351.161214JS05STAF@MIAMIU.BITNET>
Date: 17 Dec 90 21:12:14 GMT
References: <1990Dec13.011756.22888@phri.nyu.edu>
Organization: Miami University - Academic Computer Service
Lines: 19

Concerning the questions about directions for scientific computing centers,
I think the answers tend to site specific.  Some services that I
personally believe can be profitably provided include:
 
Network access, planning, evolution.  This is as much a software issue
as a hardware issue.
 
Centralized "mainframe" service.  For most purposes it is no longer
cost justifiable to provide access to computing cycles.  There is
a justification for some centralized computer services.  Access to
and maintanence of sharable software.  Backup.  Shared information
pool.  Avoidance of maintanence burdens by "end users".  Share
expensive peripherals such as vector processors, very large disk
store.  Tape drives.
 
If you provide a VAX replacement centralized computer, it seems to
me that what to provide is primarily driven by the software needs
of your client base, including both network services and application
services.