Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [RE] Focus of discussions



At 20:40 27/08/2004 -0700, David V James wrote:
>John,
>
>The reason for my stimulating email conversations was twofold:
>
>1) To answer the following requirements, some form of strawman
>   proposal seemed useful:
>     - Compatibility with the rest of 802.3
>     - Distinct identity
>     - Technical feasibility
>
>2) I get bored with precedural issues and can't wait for real
>   technical work to unfold.
>
>While the second item isn't really necessary, its much more fun,
>interesting, challenging, and a match to my engineering talents.
>
>While I suppose someone has to do it, at least within the 802
>environment, I much prefer to do the work, as opposed to working
>on proposals to do the work(:>).
>
>As to:
>>> Do we need to decide at this stage which speeds to support,
>>> whether to support jumbo frames [come to think of it, are jumbo
>>> frames allowed by 802.3 anyway?], whether to use additional pairs,
>>> etc?
>
>I would say yes, since it affects the items mentioned in (1).
>
>I think you were implying "no", but it was a bit too subtle for
>an email conversation. I really couldn't tell if that was
>a rhetorical question

It wasn't meant to be. I was looking for guidance from those who are more familiar with the way IEEE802 does things than I am. (My experience of standardisation is mostly with ISO/IEC and AES.) I suppose I'm really asking whether there's a danger that 802 might can the project if we got so interested in the engineering that we forgot to tick all the procedural boxes.



>>> [come to think of it, are jumbo frames allowed by 802.3 anyway?]
>Don't know if they are allowed by the "official" 802.3, but they
>are mandated by a significant numbers of customers, or so I hear.
>
>And, some suppliers have to be "profitable" as well as "official".
>I'm not the one to judge the scope of this support within the
>industry, but would favor jumbo frames if the commercial needs
>are widespread.

Ummm... It seems to me that if they're needed they should be in the standard and everything should support them (or at least there should be a standard way of negotiating whether they're used or not); and if there's no consensus for including them then anything claiming conformance should not require other parts of the system to support them.

However, if we think there might be a case for adding them in the future then we should certainly try not to do anything that would close off that option.



John Grant
   ___  ___  ___  ___    ___  ___  ___  ___  ___
  |   ||   ||   ||   |  |   ||   ||   ||   ||   |
  | N || i || n || e |  | T || i || l || e || s |
  |___||___||___||___|  |___||___||___||___||___|

Nine Tiles Networks Ltd, Cambridge, England
+44 1223 862599 and +44 1223 511455
http://www.ninetiles.com