Talk:100BaseVG

Contender
This article describes 100VG as a contender for Ethernet, but it was so in marketing only. During its brief existence, engineers tended to argue that 100VG was not CSMA/CD and therefore not Ethernet. Whether this is relevant in a historical context is arguable though, since CSMA/CD is all but dead at this point given the ubiquity of switched Ethernet, and the fact that at 1 Gbps and above, CSMA/CD is no longer used at all. Art Cancro 20:04, 9 August 2007 (UTC)

Why didn't it take off?
The article claims that this standard performed better than regular fast ethernet with hubs. So why wasn't it chosen? limited vendor support? network admins sticking with what they know? technical limitations? (I notice the article talks about the token staying in the hub, does that mean you could not connect two hubs together?) random chance? implementation costs? Plugwash (talk) 21:37, 12 November 2015 (UTC)


 * Ethernet switching was becoming popular at the same time. ~Kvng (talk) 15:44, 14 November 2015 (UTC)

Complex, Expensive, Overkill
Forgive me for being a fan, I was working for HP when this product was being marketed - it was a great many good things.

The chaotic thing about Ethernet at the time was you really only got something like 40% - 80% of your nominal bandwidth because of chatty protocols wasting bandwidth, thus needlessly high bandwidth attempted, and resultant collisions. Switches were not generally all that smart at the time (limited fabrics, small buffers) and 10mbit was still widely used, even though 100mbit (and even gbit fiber) were in the wild.

Anylan would have been good for folks who wanted fast token ring (like automobile assembly lines. But they ALREADY had fast token ring.

For everyone else, the smart hubs were far to expensive. Also the "cat three lines" in place in many environments were often faulty (one bad wire in eight is a deal breaker) - so many installations would need to run new cat 5 wire (and test it!) anyway. In such a case 100mbit Ethernet starts to look like a good deal, since often wire costs is a bigger issue than switch cost.

In order to connect to an existing Ethernet network, you had to have a bridge/adapter. This is actually misleading, because that was more or less true of the 10mbit to 100bit connection too - except that it was "built in" to high end switches. There were in fact some 100mbit hubs (usually SOHO sorts of things) that could not direct connect to 10mbit clients, although I suppose that didn't last long.

Some references, perhaps not well vetted:

http://www.maznets.com/tech/100vg-an.htm

http://h20564.www2.hpe.com/hpsc/doc/public/display?docId=bpe01005&lang=en-us&cc=us

http://www.cnn.com/TECH/computing/9910/12/net.failures.idg/

A good solid reference: ftp://ftp.hp.com/pub/networking/software/59636588.pdf — Preceding unsigned comment added by 199.46.251.143 (talk) 23:30, 9 June 2017 (UTC)


 * You're welcome to add to the article as long as it's sourced and not "fan material". Imho, the main drawbacks were its inherent half-duplex nature (although the hubs could've been replaced by switches with a dedicated token for each port) and it's single-vendor status ie. high price point. --Zac67 (talk) 10:21, 10 June 2017 (UTC)

ISO ratification
The article states that 100BaseVG was "ratified by the ISO in 1995." The IEEE standards site I cite in my recent modification to the article indicates that ANSI approved the standard in 1996. The IEEE standard is identified as "ANSI/IEEE 802.12-1995" which implies to me that it was likely approved by IEEE in 1995. I'm not familiar with ISO ratification vs. IEEE/ANSI ratification. Was there an actual action by ISO in relation to 100BaseVG? Or was there merely a relationship between IEEE/ANSI whereby IEEE/ANSI ratification automatically carried an implied ISO ratification? The bottom line is that I'm questioning the accuracy of the ISO ratification statement in the article but haven't been able to find documentation to resolve my concerns. LamoniDave (talk) 16:27, 31 July 2018 (UTC)


 * ISO (international), IEEE (professional) and ANSI (US industry & government) are very different bodies for standardization. It is quite possible and likely that they approved the standard at different times. 100BaseVG was developed by HP and submitted to the different bodies for approval. IEEE wasn't that prevalent then as they are today. --Zac67 (talk) 16:40, 31 July 2018 (UTC)


 * The article claims to be based on the Free On-line Dictionary of Computing and seems to match much of the information in The Telecommunications Handbook . The article and the two apparent sources discuss the standard being developed within IEEE committees (802.3 then 802.12) and IEEE documents its 1995 standard as having achieved ANSI approval in 1996.  One of the apparent sources for the Wikipedia article has its discussion on 100BaseVG as a part of its discussion of other IEEE standards.  In fact, the HP 100VG-AnyLAN Technology Guide states, "100VG-AnyLAN is a new network technology, defined by the IEEE 802.12 standard..." with no reference to it having been standardized by ISO. The HP "Technology Guide" does mention that it relies on some ISO standards and documents where "the IEEE 802.12 100VG-AnyLAN standard" fits within "the ISO Open Systems Reference Model." I find no discussion in my web searches outside this Wikipedia article and its sources that mention ISO approval. I'm inclined, based on the information I'm able to find at this time, to change the article to indicate approval by IEEE in 1995 rather than by ISO. LamoniDave (talk) 07:01, 1 August 2018 (UTC)