Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!ucla-cs!ames!ucbcad!ucbvax!WESLEYAN.BITNET!SST.D-BIGELOW%KLA.WESLYN
From: SST.D-BIGELOW%KLA.WESLYN@WESLEYAN.BITNET (Douglas Bigelow)
Newsgroups: comp.os.vms
Subject: TCP/IP terminal server access to VAXes
Message-ID: <12296607511.22.SST.D-BIGELOW@KLA.WESLYN>
Date: Wed, 22-Apr-87 15:02:28 EST
Article-I.D.: KLA.12296607511.22.SST.D-BIGELOW
Posted: Wed Apr 22 15:02:28 1987
Date-Received: Sat, 25-Apr-87 12:54:32 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 17

Wesleyan is replacing a DECSYSTEM-20 with another machine which may end up
being a VAX 8550 or 8700.  If so, we're planning to have access to the machine
be 100% through Ethernet via TCP/IP terminal servers.

We're running the CMU version of Tektronix TCP/IP.  Now, the questions:

1) Does anyone have any experience running 100 or so concurrent users over
   a TCP/IP network to a VAX?  (Via any software, not just Tektronix.)

2) Can anyone guess at the performance penalty we'd pay, compared with using
   straight serial connections via DMF-32s or something?

I know that LAT terminal servers would work much more efficiently, but we
have other non-DEC mainframes which are only accessable through TCP/IP.

Thanks for any opinions or advice you can offer.
-------