home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!uvaarpa!darwin.sura.net!paladin.american.edu!auvm!SUVM.BITNET!BIG-REQ
- Message-ID: <BIG-LAN%93012616533898@SUVM.SYR.EDU>
- Newsgroups: bit.listserv.big-lan
- Approved: NETNEWS@AUVM.AMERICAN.EDU
- Date: Tue, 26 Jan 1993 15:10:26 EST
- Sender: Campus-Size LAN Discussion Group <BIG-LAN@SUVM.BITNET>
- From: BIG-REQ@SUVM.BITNET
- Subject: BIG-LAN Digest, Volume 5, Number 4, Tuesday, January 26, 1993
- Lines: 140
-
- BIG-LAN DIGEST Tuesday, 26 January 1993 Volume 5 : Issue 4
-
- Today's Topics:
-
- running NON gel-filled UTP
- Routers as backbones
-
- Moderated by John Wobus, Syracuse University
-
- Relevant addresses:
- Internet BITNET
- Submissions: big-lan@suvm.acs.syr.edu BIG-LAN@SUVM
- Subscriptions: big-lan-request@suvm.acs.syr.edu BIG-REQ@SUVM
- LISTSERV/Archives: listserv@suvm.acs.syr.edu LISTSERV@SUVM
- Moderator: jmwobus@syr.edu JMWOBUS@SYREDU
- Anonymous ftp archives: syr.edu
-
- Note: BIG-LAN is redistributed through many mailing lists at other sites
- run by other individuals. If you subscribe(d) through such a
- "redistribution" list, you will need to remember its owner.
-
- syr.edu also has a copy of the BIG-LAN "FAQ" memo (answers to frequently
- asked questions) under the path information/big-lan/big-lan.faq
-
- BIG-LAN is also available via netnews, through newsgroup
- bit.listserv.big-lan.
-
- ----------------------------------------------------------------------
-
- Date: 25 Jan 93 17:41:47 U
- From: "Dave Clark" <dave_clark@igppqm.ucsd.edu>
- Subject: running NON gel-filled UTP
-
- running NON gel-filled UTP level 5 underground
- Experts-
- My goal is to try to design a point-to-point level 5 UTP copper and fiber
- network from approx 100 desktops spread over three new buildings to a single
- MDF in a fourth new building via new underground conduit. The runs would all
- be less than 330' from desk to MDF (ah, point to point flexibility) but would
- mean having to figure out a way to safely lay non-gel-filled level 5 UTP cable
- in an underground conduit system for no more than 250' of each cable's total
- length.
- My concerns are not only that gel-filled level 5 UTP cable doesn't exist, but
- that just the existence of water vapor - over long term - may affect the cable.
- I thought about running it in innerduct and perhaps purging the tubes with
- nitrogen, then sealing. In light of my desire to attempts to keep the design
- point-to-point, can anyone offer any insight/wisdom/an alternate solution on
- the moisture/gel-filling requirement?
- Thanks-
- Dave Clark
- dclark@ucsd.edu
- Institute of Geophysics and Planetary Physics
- Scripps Institution of Oceanography
- University of California San Diego
-
- -------------------------------------------------------
-
- Date: Tue, 26 Jan 93 14:16:59 -0500
- From: "John M. Wobus" <jmwobus@mailbox.syr.edu>
- Subject: Routers as backbones
-
- Dave Williams writes:
- >I'd like to hear from any large sites that have routers configured as
- >backbones. There's lots of marketing hype from the router companies
- >on why you "need" to do this, but very little in the way of hard info
- >on configuration, performance and pitfalls.
- >
- >Our current configuration consists of about 400 SPARC clients fed by 13 Sun
- >4/490 servers. Each server has multiple NC400 Network Co-processors (one per
- >client network), 2 IPI controllers and 6-8 IPI disks (some Saber's, some
- > Elites)
- >
- > [lots of text, charts, and ideas deleted: see original posting]
-
- Our network is more general-purpose than yours: lots of PCs and
- Macintoshes in lots of buildings in addition to workstations. We went
- to a "backbone in a router" strategy simply because it was the least
- expensive way to build a routed campus network.
-
- We do have some workstations & servers and worry about the same issues
- that are worrying you. You seem to have done more measurement and
- experimentation than we have. Our situation is a bit different: we
- have probably fewer workstations than you do and while some are grouped
- together, others are scattered all over campus among our PCs and
- Macintoshes. Also, we have a large timesharing Unix system which
- shares data and a lot of applications with the workstations.
-
- We chose to have one large server for user data (home directories) and
- one large server for applications software. Right now, both the
- servers are hanging on a single Ethernet which runs into a router. The
- timesharing system is on a different Ethernet, there are a couple of
- Ethernets that have between 5 and 20 workstations, and other
- workstations are scattered around the campus:
-
- | | | | | |
- +--Software | | |WSs |WSs |
- | server +---Timesharing |WSs | & | & | . . .
- | | system | |micros |micros |
- +--Data server | | | | |
- | | | | | |
- +----+------------------+-----------------+----+--------+--------+--------+
- | Router |
- +-------------------------------------------------------------------------+
-
- To make this work and to help address the problem of WSs scattered all
- over the place, we run no diskless nodes: temp and swap are on a local
- disk.
-
- We see the Ethernet with the two servers as the bottleneck and plan to
- add more Ethernets to the servers without turning on routing. For
- example, there might be a single Ethernet that ties together the
- Applications server and Data server with the Timesharing system. This
- would offload some traffic from the router as well as two Ethernets.
- An Ethernet could link the servers with group of co-located
- workstations in similar fashion. Another idea is that we could
- duplicate the software server for performance if necessary.
-
- A practical consideration that has faced us is whether we can use the
- dual paths for redundancy, either through the use of manual commands,
- or through automatic adjustment of routing. Since we use the
- automounter, we have had to think about how we might accomplish this
- and haven't tested a solution to our satisfaction. In any case, I am
- not a big fan of redundant connections: they are necessary in WANs,
- they might be useful and manageable on a Campus LAN for tying together
- a specific set of routers that use a good, fast-converging routing
- protocol, but a Campus LAN with an odd redundant link here and there is
- likely to create more problems than it solves.
-
- We have dreams of servers on an FDDI backbone with routers linking the
- backbone to individual Ethernets. So far, I have not been too
- impressed with FDDI.
-
- John Wobus
- Syracuse University
-
- -------------------------------------------------------
-
-
- End of BIG-LAN Digest
- *********************
-