eli@spdcc.COM (Steve Elias) (09/13/89)
u> hedrick@geneva.rutgers.edu (Charles Hedrick) writes: >We do not recommend running copper of any kind between buildings. You >can sometimes get away with it, but it's asking for trouble. Rather >than running ethernet coax, you should use fiber. consider using broadband coax as well. and if your buildings are scattered around town, note that some cable TV companies rent bandwidth on their cables. you might consider running ethernet to 802.4 bridges across long distance broadband cable in order to achieve connectivity which rivals that of microwave links. you might be able to guess whose .3 to .4 bridge i would recommend for this job. two of the three .3 to .4 bridges on the market will provide a high speed solution to inter-building connectivity using broadband cable... steve elias -- ... Steve Elias (eli@spdcc.com);6178906844;6179325598; {} /* free email to fax gateway for destinations in metro Boston area. */ /* send email and the destination fax number... */
hoey@ai.etl.army.mil (Dan Hoey) (09/18/89)
In article <4574@ursa-major.SPDCC.COM> eli@ursa-major.spdcc.COM (Steve Elias) writes: > consider using broadband coax as well.... But of course, in the words of Charles Hedrick, ``We do not recommend running copper of any kind between buildings. You can sometimes get away with it, but it's asking for trouble.'' Dan
eli@spdcc.COM (Steve Elias) (09/19/89)
In article <337@ai.etl.army.mil> hoey@ai.etl.army.mil (Dan Hoey) writes: > eli@ursa-major.spdcc.COM (Steve Elias) writes: > >> consider using broadband coax as well.... > >But of course, in the words of Charles Hedrick, ``We do not recommend running >copper of any kind between buildings. You can sometimes get away with it, but >it's asking for trouble.'' there are thousands of sites which run broadband and other coax between buildings with no problem. many of them do run computer networks across such cables, as well! many thousands more run cable TV channels across broadband cable between buildings... i'm not sure if 4 out of 5 networking cats would agree with Mr. Hedrick's recommendation. -- ... Steve Elias (eli@spdcc.com);6178906844;6179325598; {} /* free email to fax gateway for destinations in metro Boston area. */ /* send email and the destination fax number... */
kwe@bu-cs.BU.EDU (kwe@bu-it.bu.edu (Kent W. England)) (09/19/89)
In article <337@ai.etl.army.mil> hoey@ai.etl.army.mil (Dan Hoey) writes: >In article <4574@ursa-major.SPDCC.COM> eli@ursa-major.spdcc.COM writes: > >> consider using broadband coax as well.... > >But of course, in the words of Charles Hedrick, ``We do not recommend running >copper of any kind between buildings. You can sometimes get away with it, but >it's asking for trouble.'' > That generalization is too broad. "Copper" when taken in the context of Ethernet is bad news over any great distance, since there are potential ground loop and lightning induction problems that Ethernet per se does not deal with. Broadband coax is specifically designed for running between buildings and ground loops, lightning, and other electrical problems are specifically dealt with. The fact that you might run Ethernet frames on the broadband medium is of no consequence. However, given the choice of broadband or fiber for a simple little two point link, I would take the fiber as much simpler to install and maintain. You want to broadcast TV, avoid fiber and go with broadband coax. Kent England, Boston University
hedrick@geneva.rutgers.edu (Charles Hedrick) (09/19/89)
Broadband is an exception. My recommendation against copper was in response to a question about running Ethernet coax between buildings. Broadband is specifically designed for interbuilding work. I don't claim to know broadband technology in any detail, so I don't know whether it is as immune to lightning, differing ground potentials, etc., as fiber is. I'll defer to broadband experts for that assessment. My comments were intended for "normal" wires, things like Ethernet coax, RS232 cable, etc. Yes, we've had people run twisted pair between buildings. Eventually DEC got tired of replacing DZ's.
chris@mimsy.UUCP (Chris Torek) (09/19/89)
>In article <337@ai.etl.army.mil> hoey@ai.etl.army.mil (Dan Hoey) writes: >>But of course, in the words of Charles Hedrick, ``We do not recommend running >>copper of any kind between buildings. You can sometimes get away with it, but >>it's asking for trouble.'' In article <4609@ursa-major.SPDCC.COM> eli@spdcc.COM (Steve Elias) writes: >there are thousands of sites which run broadband and other coax >between buildings with no problem. many of them do run computer >networks across such cables, as well! many thousands more run >cable TV channels across broadband cable between buildings... > >i'm not sure if 4 out of 5 networking cats would agree with >Mr. Hedrick's recommendation. I think they would---certainly so if they had experience with both broadband and fiber: a) broadband is slow. You can get 10 MB/s by using three broadband channels (TV channel bandwidth is ~4.5 MB/s), as in DEC's product(s?). You can buy 80 MB/s off the shelf from Proteon. (18 TV channels anyone? :-) ) b) broadband requires maintenance. A large broadband cable plant (such as the one at the University of Maryland) requires a full-time person simply to go around and adjust head ends. (If you have ever wondered why it is that your cable TV reception improves every time the line is restored after it gets zapped by lightning, it is because the average cable TV plant does not bother adjusting their repeaters. This causes some minor picture degredation on a TV set, but similar `tuneup problems' are pretty much fatal to networking.) By and large, one pulls fiber, plugs it in, and forgets about it. Lightning strikes do not require replacing transceivers, etc. I am not up on the current costs for new installations of each. It does seem obvious (which does not make it necessarily true) that it should cost less to use existing broadband facilities than to run new fiber optic lines, at least initially (maintenance costs add up fast). But fiber is much nicer, and should be cheaper in the long run. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris
chris@mimsy.UUCP (Chris Torek) (09/19/89)
In article <19661@mimsy.UUCP> I wrote: >10 MB/s ... 4.5 MB/s ... 80 MB/s ... Aaack! I must have been suffering from an attack of `marketing inflatitus'. All of those should read `Mb/s': mega *bits* per second. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris
chimiak@umbc3.UMBC.EDU (Mr. William J. Chimiak ) (09/20/89)
Why not use FDDI? Interphase just announced their single ring interface that allows single MAC attachement. With 2 boards you get fault tolerant dual MAC access. The board delivers 100Mbps at the data link layer if I am not mistaken. With upper layer protocols invoked (XTP and the like) 20 Mbps has been realized at a node. Because the PHY and PMD layers support the full 100 Mbps, you can have some EMI resistant high speed networking with true future growth potential as it is an ANSI standard. The latest issue of EDN talks of broadband networking with FDDI. In it, they predict a $500/node cost as economies of scale come into play.
henry@utzoo.uucp (Henry Spencer) (09/20/89)
In article <2314@umbc3.UMBC.EDU> chimiak@umbc3.umbc.edu.UMBC.EDU (Mr. William J. Chimiak (MMA)) writes: >The latest issue of EDN talks of broadband networking with FDDI. In it, >they predict a $500/node cost as economies of scale come into play. Before getting too excited about this, compare with the predictions some years ago about how the price of Ethernet was going to drop as the same economies of scale got going. It has dropped... but not nearly as much as predicted. A cautious man would probably predict that FDDI prices will fall very gradually, not suddenly and massively. For the immediate future, it's likely to remain a "premium" networking technology, used only where Ethernet is not good enough. That's going to limit those "economies of scale" quite a bit. -- "Where is D.D. Harriman now, | Henry Spencer at U of Toronto Zoology when we really *need* him?" | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
bob@rel.mi.org (Bob Leffler) (09/21/89)
In article <19661@mimsy.UUCP>, chris@mimsy.UUCP (Chris Torek) writes: > >In article <337@ai.etl.army.mil> hoey@ai.etl.army.mil (Dan Hoey) writes: > In article <4609@ursa-major.SPDCC.COM> eli@spdcc.COM (Steve Elias) writes: > >i'm not sure if 4 out of 5 networking cats would agree with > >Mr. Hedrick's recommendation. > I think they would---certainly so if they had experience with both > broadband and fiber: > > I am not up on the current costs for new installations of each. It > does seem obvious (which does not make it necessarily true) that it > should cost less to use existing broadband facilities than to run new > fiber optic lines, at least initially (maintenance costs add up fast). > But fiber is much nicer, and should be cheaper in the long run. I second Chris's opinion. I support about 3000 users on a combination of fiber, broadband, and thick ethernet. Fiber has proven to be very cost effective, both in installation and maintenance. Because of this, I'm in the process of moving the users that are on the braodband and thick ethernet plants to our new fiber plant. bob -- Bob Leffler - Electronic Data Systems, Financial Information Services Division 3044 West Grand Blvd., Room 11-101, Detroit, MI 48202 (313) 556-4474 bob@rel.mi.org or {uunet!edsews, rutgers, sharkey}!rel!bob Opinions expressed may not be those of my employer.
gary@dgcad.SV.DG.COM (Gary Bridgewater) (09/21/89)
In article <4609@ursa-major.SPDCC.COM> eli@ursa-major.spdcc.COM (Steve Elias) writes: >In article <337@ai.etl.army.mil> hoey@ai.etl.army.mil (Dan Hoey) writes: >> eli@ursa-major.spdcc.COM (Steve Elias) writes: >>> consider using broadband coax as well.... >>But of course, in the words of Charles Hedrick, ``We do not recommend running >>copper of any kind between buildings. You can sometimes get away with it, but >>it's asking for trouble.'' > i'm not sure if 4 out of 5 networking cats would agree with Yes, but that fifth story is liable to put the fear of god into anyone who hears it. When I was an SE in Denver - a well known place to find lightning - I saw the innards of two different systems which had long wires between buildings or floors that had been hit by lightning. The boards were all charred, most of the chips had melted, some had popped off the board. The power supplies were fused and all the plastic was coated with ash. The bigger system was down for 10 days while the FEs rebuilt it from the ground up. That systems was well inside a building as was the wire - so naturally they dropped their lightning insurance. It seems the lightning followed a crack in the concrete or possibly vent pipes - there was too much damage to be sure. If you do run wires in a lightning prone area consider trenching. -- Gary Bridgewater, Data General Corp., Sunnyvale Ca. gary@sv4.ceo.sv.dg.com or {amdahl,aeras,amdcad,mas1,matra3}!dgcad.SV.DG.COM!gary No good deed goes unpunished.
goodloe@b11.ingr.com (Tony Goodloe) (09/21/89)
In article <2314@umbc3.UMBC.EDU>, chimiak@umbc3.UMBC.EDU (Mr. William J. Chimiak ) writes: > .. EDN talks of broadband networking with FDDI. Broadband with FDDI? I don't follow. Someone enlighten me (nicely). tony goodloe
hermann@acadch.UUCP (Hermann Buergi) (09/21/89)
Hi out there,
we're looking for a laser- or microwave bridge to join the ethernets of two
buildings. The distance between them is aproximately 1.5 miles. Does anyone
run such a configuration ?
Any pointers will help.
Thanks in advance
Hermann
--
Hermann Buergi, Autdoesk Switzerland, hermann@acadch.commorgan@Jessica.stanford.edu (RL "Bob" Morgan) (09/22/89)
> A cautious man would probably predict that FDDI prices will fall very > gradually, not suddenly and massively. For the immediate future, it's > likely to remain a "premium" networking technology, used only where > Ethernet is not good enough. Moreover, I'll bet that the most common application on these fancy FDDI-attached workstations will still be Telnet . . . *8^)* - RL "Bob"
smb@ulysses.homer.nj.att.com (Steven M. Bellovin) (09/23/89)
In article <1130@svx.SV.DG.COM>, gary@dgcad.SV.DG.COM (Gary Bridgewater) writes: [lightning horror stories deleted] > If you do run wires in a lightning prone area consider trenching. Lightning is quirky stuff, and damned hard to defend against. I've seen situations where long runs of cable within the same building have ``attracted'' the attention of lighting bolts, frying the equipment on either end of RS-232 lines. Based on my own experiences, I don't think buried cables -- I assume that that's what you mean by ``trenching'' -- help that much. When I was in Chapel Hill, we regularly lost some gear connected by cables that ran under the street. We tried all sorts of protectors, too; nothing seemed to help. The problem was that in the time (and at the voltage) the lightning protection devices took to act, it was already too late -- the chips are just too sensitive. We eventually resorted to optoisolators; they'd get fried, too, but it was easier and cheaper just to replace them.... For that matter, during one particularly severe storm, my TV was fried via the cable TV input jack. Of course, that strike was close enough that it blew out a couple of light bulbs and even popped one of my apartment's circuit breakers -- I'm not sure anything but a Faraday cage would have helped against that one.
karn@ka9q.bellcore.com (Phil Karn) (09/23/89)
In article <12204@ulysses.homer.nj.att.com> smb@ulysses.homer.nj.att.com (Steven M. Bellovin) writes: >Lightning is quirky stuff, and damned hard to defend against. Agreed. Its effects are a little less mysterious, though, if you analyze everything in terms of ground loops. When you're talking about peak currents of 10,000 to 50,000 amperes, there's just no such thing as "common ground potential". I've read reports of lightning hits that killed only certain people and animals apparently at random out of a much larger group that was standing or lying near the hit. When the case was analyzed, however, it was discovered that the fatalities were generally among those people or animals who were lying or standing either towards or away from the point the lightning struck. This exposed them to larger voltage gradients along the ground. Those lying or standing perpendicular to the line to the strike point had a much greater chance of surviving. The point is that IR drops are everywhere during a lightning hit, especially in the ground itself (which is not a particularly good conductor, even when wet). It's impossible to get rid of them altogether with even the heaviest of ground conductors. The best you can do is to use single-point grounding schemes so that everything rides up and down together during the strike, but this takes considerable attention to detail even in small systems. It's practically impossible between large buildings. If at all possible, I *strongly* recommend fiber between buildings. Phil
chimiak@umbc3.UMBC.EDU (Mr. William J. Chimiak ) (09/27/89)
> Broadband with FDDI? I don't follow. Someone enlighten me (nicely). > > tony goodloe I am sorry for perhaps being misleading, but it is my impression that the nature of the question is the nagging problem that workstations now have the I/O and processing speeds which blow past the capabilities of ETHERNET. Folks with image processing applications and huge file transfers are irritated with the LAN/CAN/WAN bottleneck. One of the major problems is the media access which roughly resides in the Data Link layer (layer 2) of the OSI model. FDDI is capable of delivering 100 Mbps at the MAC layer. Besides being that fast, FDDI has the potential of delivering real time performance - but this must be said with a caveat. Real time performance correctly implies an Application Layer capability. This is the final layer of the OSI model. To deliver true real time performance obviously implies that the layers three though six deliver the same real time performance. NIST is studyinging realtime performance of Layer 4 while folks like Gregg Chesson are developing the eXpress Transport Protocol (XTP) to at least give workstations the capability of operating on Gbps LANs.