info-vax@ucbvax.ARPA (10/17/84)
From: mark thompson <THOMPSON@USC-ECLC.ARPA> Has anybody hooked a T1 line up to their vax? Has anybody got a way to join two ethernets that are seperated by about a mile? -mark <thompson@usc-eclc.arpa> -------
info-vax@ucbvax.ARPA (10/18/84)
From: God <root%bostonu.csnet@csnet-relay.arpa> From unix-wizards-request@BRL-TGR.CSNET Wed Oct 17 02:47:01 1984 Date: Tue 16 Oct 84 16:18:50-PDT From: mark thompson <THOMPSON@USC-ECLC.CSNET> Subject: Fast communications To: unix-wizards@BRL-VGR.CSNET, info-vax@SRI-KL.CSNET Organization: University of Southern Calif. Computing Services Phone: (213) 743-4800 Via: csnet-relay; 17 Oct 84 2:23-EDT Has anybody hooked a T1 line up to their vax? Has anybody got a way to join two ethernets that are seperated by about a mile? -mark <thompson@usc-eclc.arpa> ------- Boston University is currently running 4.2bsd TCP/IP over our Ungermann/Bass Broadband (also to VMS/EUNICE/TCP-IP.) We are getting ethernet-like speeds (FTP of TERMCAP @>200Kb/s eg.) We are using a locally developed driver based on the 4.2 un driver which was designed for the old 3Mb baseband. The initial installation of a broadband system ain't cheap and is really an organization-wide decision. However, once done + our software we are able to do high speed service between any two points on campus (miles.) In addition, we are installing things like microwave bridges to tie together our broadband with one at our medical campus (T1'ish) across town. Maybe U/B has a 'starter' system? -Barry Shein
info-vax@ucbvax.ARPA (10/18/84)
From: Jonathan Goodman <Goodman@YALE.ARPA>
Has anybody got a way to join two ethernets that are seperated by
about a mile?
-mark <thompson@usc-eclc.arpa>
-------
We accomplished this using a GE Gemlink which is buffered on both ends
by 11/34's running a hacked up version of the old Minits. If one can
tolerate the occasional flakiness of rf communications, it's a reasonable
solution. If you need 100% reliability it's not the way to go. Cornell
has had experience with this product too. --Jon
-------info-vax@ucbvax.ARPA (10/18/84)
From: Richard Garland <OC.GARLAND%CU20B@COLUMBIA.ARPA> -- Joining 2 ethernets separated by a mile: If you can squeez the distance to a kilometer (1000M) then you can get an ethernet repeater with an optical fiber link between 2 halves of the repeater. DEC sells such a product and has shown it at the last several DECUS symposia. Perhaps you can bend the specs a little and run it over a longer fiber. The name of this box is the DEREP-RA. It consists of 2 half repeaters. You also buy the fiber and 2 tranceivers to hang off of each ethernet. There may be similar products from others - I think Ungermann-Bass has something similar. This scheme has the advantage of a single E-net. I'm not sure how the cost compares to other schemes such as T1 lines hooked to PDP-11's etc. It may also have a speed advantage. Rg -------
info-vax@ucbvax.ARPA (10/18/84)
From: Marty Sasaki <sasaki%harvard.arpa@csnet-relay.arpa>
Is the stuff you are doing point-to-point but using the cable as the wiring?
Harvard has a broadband and we are using Sytek to do terminal/host
communication and are looking at high speed stuff. We have thought about
buying some modems and doing point-to-point, but that will severely
limit the number of nodes.
Marty Sasaki
Harvard University Science Center
sasaki@harvard.{arpa,uucp}info-vax@ucbvax.ARPA (10/19/84)
From: decvax!genrad!bolton!garry@BERKELEY (Garry Baer) MICROWAVE Not cheap but it will get the job done!!! Garry Baer vax!decwrl!ucbvax!info-vax@Berkeley Subject: Re: Fast communications In-Reply-To: your article <2608@ucbvax.ARPA> MICROWAVE Not cheap but it will get the job done!!! Garry Baer
info-vax@ucbvax.ARPA (10/19/84)
From: medin@ucbarpa.BERKELEY (Milo Medin) also, vitalink in mountain view makes such a beat and its available right now. Milo
info-vax@ucbvax.ARPA (10/19/84)
From: Lee Moore <lee@rochester.arpa>
If you are using an inter-networking protocol, such as TCP/IP or XNS,
you could put a gateway on each ethernet and use a long-haul network
to talk between them. From my ethernet in Rochester, NY I can talk over
the Arpanet to ethernets at Stanford, Berkeley, CMU, etc. We also use
4.2BSD Unix to talk over a point-to-point link to machine(s) at another
campus. Currently this link is rs-232 but 4.2 has drivers for things
like DMR-11s which approach T1.
=lee
Internet: lee@rochester.arpa
UUCP: {decvax, allegra, seismo}!rochester!lee
Phone: [USA] (716) 275-7747, -5671
Physical: 43 01' 40'' N, 77 37' 49'' Winfo-vax@ucbvax.ARPA (10/19/84)
From: medin@ucbarpa.BERKELEY (Milo Medin) The advantage of the vitalink system is that it forwards ethernet packets, regardlless of the protocol they are being used in. You can run IP and DDCMP over the same bridge if you want. They did it the right way. Milo
info-vax@ucbvax.ARPA (10/21/84)
From: Murray.pa@XEROX.ARPA "They did it the right way." Oh, yea? Consider what happens if you connect 3 Ethernets with 3 bridges so you have a backup path if any one goes down.
info-vax@ucbvax.ARPA (10/21/84)
From: medin@ucbarpa.BERKELEY (Milo Medin) Granted you cant have a loop in the path, but that cluster is logically one ethernet. By using a host with 2 interfaces as a gateway, you can get around that. The key point here is that it doesnt force you into any protocol which a gateway does. Milo
%SRI-KL.ARPA:medin%ucbarpa%Berkeley@randvax.UUCP (10/28/84)
From: medin%ucbarpa@Berkeley (Milo Medin) Granted you cant have a loop in the path, but that cluster is logically one ethernet. By using a host with 2 interfaces as a gateway, you can get around that. The key point here is that it doesnt force you into any protocol which a gateway does. Milo