Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site ucsfcca.UUCP
Path: utzoo!linus!decvax!decwrl!ucbvax!ucsfcgl!ucsfcca!dick
From: dick@ucsfcca.UUCP (Dick Karpinski)
Newsgroups: net.lan
Subject: Re: Ether-Ether machine
Message-ID: <407@ucsfcca.UUCP>
Date: Fri, 13-Sep-85 21:18:07 EDT
Article-I.D.: ucsfcca.407
Posted: Fri Sep 13 21:18:07 1985
Date-Received: Sun, 15-Sep-85 05:17:20 EDT
References: <515@lasspvax.UUCP>
Reply-To: dick@ucsfcca.UUCP (Dick Karpinski)
Organization: UCSF Computer Center
Lines: 46

In article <515@lasspvax.UUCP> conrad@lasspvax.UUCP (Conrad Cady) writes:
>
>     Is there an Ethernet to Ethernet dedicated packet-passing machine that
>anyone know of?  Do you know where it can be purchased, or the price?
>
>     Here's our situation.  We want to attach ourselves to a local Ethernet,
>and were allowed one connection, but we'd like to hook several workstations
>and devices into it.  We've decided that we'd like our own Ethernet, with a
>gateway to the other one.  Rather than having our 750 act as the gateway,
>we'd like to explore the possibility of using some dedicated box as the
>gateway.  
>
There are lots of possibilities:
  1. A DEC DELNI can be used as a multi-station transciever.  In fact,
     if there were no other (remote) machines, I think you wouldn't
     need any Ethernet cable at all.  I believe someone else makes a
     similar device, perhaps TCL (415) 657-3800.
  2. A repeater makes only one connection, but it must be figured in
     the maximum length calculations and of course all traffic is
     seen on both Ethernet wires.  DEC and Ungermann/Bass (408) 496-
     0111 and InterLan (617) 692-3900, among others, supply such.
  3. A buffered repeater makes only one connection and does not
     affect max length calculations.  All traffic is, however, still
     seen on both wires.  I believe U/B makes one.
  4. Filtered repeaters would only pass traffic intended for the
     other wire, but I know of none.
  5. Bridges and gateways pass some or all of the traffic but also
     impose other constraints:  
     a. Only one protocol (DECnet, TCP/IP, and XNS are all popular)
     b. Limited throughput (VitaLink's TransLan handles 224kbps max)
     c. Must be explicitly addressed (only the TransLan listens to
        all traffic, passing as appropriate, as I understand it.
        This is the elusive MAC (media access control?) level bridge.)
     d. High cost compared to repeaters, even buffered ones.

Is that enough more than you wanted to know?

Dick

ps.  PLEASE flame me if I got any of that wrong!

-- 
Dick Karpinski    Manager of Unix Services, UCSF Computer Center
UUCP: ...!ucbvax!ucsfcgl!cca.ucsf!dick   (415) 666-4529 (12-7)
BITNET: dick@ucsfcca   Compuserve: 70215,1277  Telemail: RKarpinski
USPS: U-76 UCSF, San Francisco, CA 94143