BIG-MOD@SUVM (02/14/90)
BIG-LAN DIGEST Tuesday, 13 February 1990 Volume 2 : Issue 11
Today's Topics:
"NSFnet router" - IP/routers
LAN magazine
PC Compatibles?
If you have advice on measuring the traffic on an ISN
Alan Watt asked about experiences with Twisted Pair Enet components.
Internet Engineering Task Force meetings.
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: listserv@suvm.acs.syr.edu LISTSERV@SUVM
Moderator: jmwobus@suvm.acs.syr.edu JMWOBUS@SUVM
----------------------------------------------------------------------
Date: Mon, 5 Feb 90 13:23:38 EST
From: magill@eniac.seas.upenn.edu (Network Planning)
Subject: "NSFnet router" - IP/routers
> a friend of mine, from the brazilian institution LNCC (National Laboratory of
> Scientific Computing) requested me some information about a NSFNet router
> made by PROTEON. Would someone please give me some information about this
> equipment ? I mean something about reliability, performance, and some other
> information that could help him in his buying decision ?
There are two principal IP routers on the market today.
Proteon and cisco. (cisco is really spelled lower case it's an
"extraction" from another word (SanFran....)
We happen to have both - Proteon on our PREPNET connection and
cisco on our JVNCnet connection.
They both work, do what they are supposed to do, but we prefer
the cisco box. It has a number of features (like able to multiply
bandwidth by the number of lines used to the same destination,
compatible terminal servers, etc) which make us feel it is a superior
box. I don't know what the price comparison happens to be, but
depending upon the number of routers needed the cost of the cisco
drops dramatically. We presently have 6 routers in one chasis.
The cisco also supports fun things like FDDI, DECnet, AppleTalk,
has an assortemnt of interfaces to fun things like X.25, Ethernet, etc.
I don't have the actual stats, but I believe that the actual
throughput (packets per second) is significantly higher on the
cisco equipment. This can be significant if one is using multiple
circuits, but since the Proteon can't, the difference is somewhat
academic - the Proteon does route at T1 speeds. (All our external
connections are T1 at this time.
William H. Magill Manager, PennNet Operations Planning
Data Communications and Computing Services (DCCS) University of Pennsylvania
Internet: magill@dccs.upenn.edu magill@eniac.seas.upenn.edu
magill@upenn.edu
-------------------------------------------------------
Date: Mon, 5 Feb 90 14:51 GMT
From: "Robert Demaine (Computing Service)" <RLD1@VAXA.YORK.AC.UK>
Subject: LAN magazine
Has anyone had difficulty obtaining copies of LAN Magazine?
We took out a year's subscription last April after it had been mentioned in
this
list and received regular copies until September. Then they just stopped
coming. I've tried contacting the publishers by letter and by fax but
without success.
I found it a useful publication and would like to continue receiving it.
Robert Demaine
University of York Computing Service
York, England
-------------------------------------------------------
Date: Mon, 05 Feb 90 18:16 CST
From: Bill McGown <CFWPM@ECNCDC>
Subject: PC Compatibles?
Eastern Illinois University has had in place for a number of
years a policy that has restricted the purchase of microcomputers to
two vendors; IBM and Zenith. An administrative decision has been
made that it might be beneficial if EIU were to have the choice
of one more vendor of an "IBM compatible" computer. The main
requirement of this third vendor's computer is that it be
compatible with our token ring network. This network is based on
the IBM Token-ring network technology. Our local rings are
running across IBM Type 1 shielded twisted pair at 4 MBits, with
a backbone ring of optical fiber running at 16 MBits. We use IBM
Token-ring Adapter IIs and /A cards for connection, off-the-shelf
IBM software, and IBM PCs, PS/2s and Zenith PCs. All of our
network services are based on IBM architectures, and we use
NETBIOS as our communications protocol.
We will need a vendor that has a "full line" of machines,
varying from low end XT types to 386 33mhz. server class
machines.
If any reader has had experience with "clone"
machines in a similar environment I would like to know the brands
of machines and how they have performed?
I will summarize responses if there is sufficient
interest and post then to the list. Pardon the cross posting but
this request is being posted to both INFO-IBMPC and BIG-LAN.
- -----------------------------------------------------------------------------
Bill (W. P.) McGown BITNET: CFWPM@ECNCDC
Psychology Dept.
Eastern Ill. Univ.
Charleston, IL
61920
- ------------------------------------------------------------------------------
-------------------------------------------------------
Date: Thu, 8 Feb 90 08:55:01 CST
From: "Lynn A. Christiansen" <J1.LAC@ISUMVS.IASTATE.EDU>
Subject: If you have advice on measuring the traffic on an ISN
If you have advice on measuring the traffic on an ISN
backbone, I would appreciate hearing from you.
We have five packet controllers fully interconnected
on campus and would like to monitor the traffic.
Since the number of ethernet sublans is steadily
increasing, we would like to see how many EBIMs
the ISN can handle before we have traffic problems.
Thank you ahead of time.
Lynn A. Christiansen
371F Durham Center
Iowa State University
Ames, Iowa 50011
-------------------------------------------------------
Date: 11 Feb 90 14:16:00 EST
From: "DAVE DOROSZ" <dorosz@gw2.hanscom.af.mil>
Subject: Alan Watt asked about experiences with Twisted Pair Enet
components.
Alan Watt asked about experiences with Twisted Pair Enet components.
We have just started to install Synoptics boxes here and they seem
to be fairly easy to install. I can't comment on the reliability
of the Synoptics stuff, although a question I posted to this list
about twisted pair brought forth many favorable responses about
Synoptics hardware in general.
The Mitre Corp. has experimented with ODS equipment ( the fiber stuff) and
found
it to be quite unreliable. The basic problem is that the fiber transceiver
was poorly designed and could not disapate the heat generated by some
of the electronics in the box. I wonder if others who have tried ODS
hardware have found this to be a problem ?
DAVE DOROSZ
Hanscom AFB.
DOROSZ@gw2.hanscom.af.mil
-------------------------------------------------------
Date: Tue, 13 Feb 90 14:18:34 EST
From: "John M. Wobus" <JMWOBUS@SUVM>
Subject: Internet Engineering Task Force meetings.
I attended the IETF meetings just recently in Tallahassee, basically
just listening for news. Here is some news of interest to those
managing campus-sized TCP/IP-based networks:
(1) The Internet powers-that-be have been pressured to facilitate the adoption
of a suitable replacement for RIP. Two candidates were discussed
but one isn't yet written down (Dual IS-IS; the ISO routing protocol
extended to support TCP/IP) and the other has not been tested much
in the field (OSPF). No recommendation was made to adopt either one,
but the recommendation was made to consider such things again after
OSPF has been tested in Suranet. In the past, router vendors
(other than the one that developed OSPF) have been reluctant to promise to
implement it, citing the fact that it is not a standard. I think
some router vendors may be changing their mind, having seen some
"handwriting on the wall".
(2) A new RFC outlining the requirements to be an "internet router" has
been started and a 12-month schedule has been laid out for its
completion. This would be a companion to the recently completed
requirements for an "internet host".
(3) Other work that is going on: development a standard way for hosts
to be configured through the LAN; development of ways for hosts to
find out the address of a gateway they can use to get off their own
network; development of ways that a host can discover how large
a packet it can send without it being fragmented; continuing work
on MIB II for SNMP.
(4) I would say there is only a small amount of effort towards the
development and promotion of additional "standard" application
protocols (beyond TELNET, SMPT, & FTP). There is a new working
group to look at adopting a standard printing protocol (for
example, the LPR/LPD protocol) and probably some other working
groups, but the IETF seems to consist mostly of people interested
in lower level networking problems as well as problems dealing with
the Internet.
(5) There is some effort towards more "purely informational" RFCs.
A new RFC listing "Network Operations Center Tools"
is very close to being made official. Other RFCs in varying
stages of development are a bibliography, a glossary, and a
compendium of commonly asked questions along with their answers.
John Wobus
Syracuse University
-------------------------------------------------------
End of BIG-LAN Digest
*********************