Date: 10 Jun 96 09:14:37 EDT From: John H Boal <100437.2500@CompuServe.COM> To: JoAnn Larson Cc: "802.1 Mailer" Subject: P802.1p/D3 Ballot SUBJECT: P802.1p/D3 - Traffic Class and Dynamic Multicast Filtering Services in Bridged LANs _XX__ I disapprove for the following reasons. John H Boal ---------------------------------------------------------------------- 1 The scope and definition of 802.1p while realistically conceived to deal with a required enhancement of established 802.1 bridging standards has been overtaken by events not only within but also outside 802. This assertion appears to have acquired some general credence at the recent Wakefield interim meeting that I was unable to attend. Specifically there has been a recognition that "p" and now "q" have intended feature sets that abstractly display significant commonalties of purpose. The potential need for more than two levels of priority - "normal" and "expedited" - has been recognised. The potential for introducing priority as another (abstract) vLAN addressing space has also been noted. I therefore dis-approve D3. I have hesitation in noting that my observations above reflect the multicast commentaries of several other voting members. 2 Comment not directed to 802.1p/D3 specifically but relevant in my view to further work thereon. I have tried to articulate the need for a more structured and top-down approach to the work of 802 in general and 802.1 in particular. Thus far I do not seem to have enjoyed much success. I note other member's (rejecting) comments allude to the need for more levels of priority in 802.1p - and the inclusion of such capabilities within "p" and/or "q". Further, I see that the reassignment of 3 bits out of 16 bits of "established" vLAN addressing is now "controversial". In La Jolla, after there had been a unanimous vote for a "vLAN architecture", I was the pariah that spoiled the possibility of a subsequently unanimous vote for fixing a specific allocation for a vLAN "address encoding" structure. Now the need for priority is causing some members anguish! I know this is NOT the place to dis-approve "q" activities. BUT, in the absence of a more considered 802.1 strategy, I am using this platform to record (with only a little hindsight) my dis-approval of the disordered way in which both 802.1p (and "q") have set out to deal with very limited parts of much larger set of systematic problems. The IETF have similar problems but THEY are approaching them FROM a Quality of Service viewpoint. Their past success is still not a license for them to succeed in new multi-service networks. However I applaud the IETF approach, as rather more {scientific} and less wantonly {commercial}. For their trouble these "academics" might actually enjoy some commercial success yet again! John H Boal Racal Research Limited Worton Drive READING, Berkshire RG2 0SB United Kingdom