phil@pepsi.amd.com (Phil Ngai) (04/10/90)
In article <76700190@p.cs.uiuc.edu> gillies@p.cs.uiuc.edu writes: |Five years ago I heard David Redell comment on Ethernet |standardization. Dr. Redell represented Xerox in the |Xerox/Intel/DEC/IEEE 802 standardization process, in 1979. | |He complained that DEC was unreasonable in demanding 10Mb/sec |performance, when Xerox's 3Mb/sec performance would have been |perfectly adequate. | |Who was right? Today, PC's (Compaq systempro) are hitting the 10Mb/s |wall. How far should a standard push the state-of-the-art, given the |costs of doing so, the expected lifetime, and the ability to make a |new standard later? Xerox's system was based on CATV components. The media cost was extremely low. I don't know how long it took DEC's media to reach the point of diminishing returns in terms of economies of scale but certainly at my company we didn't really start to use Ethernet until 1985 and at least part of that was due to cost. DEC made things more expensive in terms of demanding higher speed *and* longer range. This is why the yellow coax is so thick and expensive. It seems to me that repeaters and bridges are a much more attractive solution. The thin-net stuff seems reasonable enough, however. Even if Compaq systempros are reaching a limit, that doesn't mean every PC needs FDDI. Multiple thin-nets to systempros seems like a good solution, especially if you are using twisted pair Ethernet where you effectively have multiple nets anyway, as long as your systempro can be close to the wire closet. -- Phil Ngai, phil@amd.com {uunet,decwrl,ucbvax}!amdcad!phil The War on Drugs is the modern day Inquisition.
henry@utzoo.uucp (Henry Spencer) (04/11/90)
In article <76700190@p.cs.uiuc.edu> gillies@p.cs.uiuc.edu writes: >He complained that DEC was unreasonable in demanding 10Mb/sec >performance, when Xerox's 3Mb/sec performance would have been >perfectly adequate. > >Who was right? Today, PC's (Compaq systempro) are hitting the 10Mb/s >wall... Suns have been able to drive an Ethernet pretty much flat-out for years now. However, they seldom do it for very long in practice. It's kind of hard to say who was right. Faster is better, other things being equal... but they aren't equal. The 10Mb/s rate was just fast enough to drastically boost the price of an Ethernet interface compared to the old 3Mb/s ones. If the standard had gone with 3Mb/s, we might (repeat, might) by now have somewhat slower Ethernets so widespread that RS232 could be dying out. Certainly local networks would have gotten started earlier and be far more widespread, given the lower costs. Most of Ethernet's current competitors would have been stillborn, with all that means in interoperability and standardization. I tend to suspect it would have been worth it. >What is the remaining lifetime of 10Mb/s ethernet? That's like asking what's the remaining lifetime of RS232. Ethernet may not be the front-line networking for the high-priced hardware for too much longer, but it will be the standard for non-performance-critical local networking for a long, long time. It's too universal to die any time soon. -- With features like this, | Henry Spencer at U of Toronto Zoology who needs bugs? | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (04/11/90)
In article <1990Apr10.195000.7522@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: | >What is the remaining lifetime of 10Mb/s ethernet? | | That's like asking what's the remaining lifetime of RS232. Ethernet | may not be the front-line networking for the high-priced hardware for | too much longer, but it will be the standard for non-performance-critical | local networking for a long, long time. It's too universal to die any | time soon. And existing enet installations can get better performance by use of a faster (usually optical) backbone. This enables any two stations to get about the same performance they would if connected by an unloaded enet. There are some application which really need to connect two machines faster than enet can run unloaded, but for many installations the problem is total load, and that can be handled. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) "Stupidity, like virtue, is its own reward" -me
pcg@aber-cs.UUCP (Piercarlo Grandi) (04/12/90)
In article <29801@amdcad.AMD.COM> phil@pepsi.AMD.COM (Phil Ngai) writes: In article <76700190@p.cs.uiuc.edu> gillies@p.cs.uiuc.edu writes: |He complained that DEC was unreasonable in demanding 10Mb/sec |performance, when Xerox's 3Mb/sec performance would have been |perfectly adequate. On this I thoroughly agree. The Xerox people were very clever guys, and they were "right". |Who was right? Today, PC's (Compaq systempro) are hitting the 10Mb/s |wall. Only because they use the wire as if it were an IO bus or even worse a memory bus. This is ridiculous, just as having dozens of diskless with a wire (which has a nonlinearly worsening response to increased load!) between them and their remote discs. I have been told by a guy (and a very bright and competent one) that for a large 80K fortran compile he run it on a local fast CPU and a remote large memory CPU as disc buffer cache because so he could have the best of both worlds; fast CPU here and large memory there. The wire was not busy otherwise. This is literally using an Ethernet as a 5ms. latency memory bus. It may be advantageous at times, but in the general case... |How far should a standard push the state-of-the-art, given the |costs of doing so, the expected lifetime, and the ability to make a |new standard later? Xerox's system was based on CATV components. The media cost was extremely low. I don't know how long it took DEC's media to reach the point of diminishing returns in terms of economies of scale but certainly at my company we didn't really start to use Ethernet until 1985 and at least part of that was due to cost. It is not just this. The Xerox guys designed Ethernet as a way of linking essentially autonomous workstations together and with wider area services, provided by servers or gateways. They never did envision the absurdity of using it as system bus for porrly engineered loosely coupled multiprocessors, or as IO bus or memory bus. They saw the world as a constellation of small Ethernet segments, with strong, *strong* locality of access, first between a workstation and its _local_ discs, and then between a workstation and its rarely accessed peers on the same wire, and then between a workstation and an even less frequently accessed server or gateway on some other wire. Given such a perspective, the performance profile of a CSMA/CD protocol are not a problem, because you expect the wire to be rarely used, even if in those cases you want the full bandwidth. Metcalfe himself hints strongly at this in his interview in a recent issue of UnixWorld. -- Piercarlo "Peter" Grandi | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk Dept of CS, UCW Aberystwyth | UUCP: ...!mcvax!ukc!aber-cs!pcg Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk
rbg@rbg.Sun.COM (Robert Garner) (04/12/90)
> In article <29801@amdcad.AMD.COM> phil@pepsi.AMD.COM (Phil Ngai) writes: > In article <76700190@p.cs.uiuc.edu> gillies@p.cs.uiuc.edu writes: > > |He complained that DEC was unreasonable in demanding 10Mb/sec > |performance, when Xerox's 3Mb/sec performance would have been > |perfectly adequate. > > On this I thoroughly agree. The Xerox people were very clever guys, and > they were "right". I just picked up on this conversation. I worked under Metcalfe in Xerox's System Development Division (SDD) in 1978 on the first 10Mb/s Ethernet controller hardware, an evolution of Boggs' Alto hardware. Ron Crane, now also at 3Com, did the first analog engineering analysis for the 10Mb Ethernet. Long before DEC entered the scene, we were trying for 20Mb/s on what was internally called the "XeroxWire." One reason we dropped back to 10Mb/s was the availability of a Fairchild CRC chip that ran at 10 MHz! Random STTL in the boards for the Dolphin (D0) and Dandelion (Star) workstation comprised just too many parts for the CRC. (Keep in mind that the Dandelion's Ethernet controller hardware was just 90 SSI chips--not bad for 1978.) Anyways, based on electrical considerations--primarily the attenuation of signals and the noise introduced by taps-- 5 MHz would have been a better speed. However, the idea was that 10 MHz looked doable, so we went for it. (The kludgery of the "put-your-transceiver-here marks" on the cable came latter.) On a different subject, since the whole world (or at least Japan, Germany, and Fance) may be going to 64Kb/s ISDN in several years, this could make 10Mb/s Ethernet look fast. - Robert Garner p.s. I assume everyone knows that the 3Mb-Ethernet was actually 2.94 Mb/s, which was half the Alto's clock rate, which was an integral submultiple of the display's line rate...? ARPA: garner@sun.com UUCP: {ucbvax,decvax,allegra,decwrl,cbosgd,ihnp4,seismo}!sun!garner Phone: (415)336-1708 ISDN: ???
wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) (04/12/90)
Can somebody tell me what the theoretical maximum rate that you can push data down an RG-58 (or even the 75 ohm RG-59) cable. Does Ethernet come close to this limit? Dean Wallwey
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (04/13/90)
In article <19596@boulder.Colorado.EDU> wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM) writes: | Can somebody tell me what the theoretical maximum rate that you can push | data down an RG-58 (or even the 75 ohm RG-59) cable. I have no idea what the LIMIT is, but this stuff is used to distribute TV and FM radio, so it's > 100 MHz limit, since that's the middle of the FM band. We're running 10MHz ethernet over twisted pair these days, so decent cable has lots of headroom. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) "Stupidity, like virtue, is its own reward" -me