FC-BB January 20, 2000 FC-BB Minutes

Attendees:

Steve Wilson (Brocade)       Murali Rajagopal (Gadzoox)
Mike O'Donnell (McData)      Jack Harwood (EMC)
Craig Carlson (Ancor)        Steve Ainsworth (Ancor)
Eric Peterson (Troika)       Ken Hirata (Vixel)
Carl Zeitler (Compaq)        Bill Collette (CNT)
Brett Ketchum (CNT)

The only item on the agenda was to review Rev 4.4 of the 
FC-BB specification. Due to most changes occurring in the clause that
describes flow control, we only reviewed that clause. We will review the
remainder of the document at the next FC-BB meeting in February.

Here are the main discussion items from the review:

It was determined that a state machine depicting the flow control states,
particularly initialization would be useful. Everyone deemed this as a
good idea. 

We also discussed whether certain unnumbered 
messages should be included. Murali stated that they were specified because
most implementations of LAPB would include them. Brett suggested that we
should entertain specifying a SNAP header. We also discussed whether class F
frames should be given a higher priority than other FC frames. Should they
be subject to flow control at all? Should this be in standards or is it 
implementation dependent? All will investigate.

McData brought up whether FS_TOV needs to be different for WAN attached
swithes than locally attached switches. We discussed whether this is an
FC-SW2 issue or an FC-BB issue. Solutions spanned from making the process
of determining FS_TOV dynamic to keeping the value as is. For now it looks
like this is an FC-SW2 issue and the switch vendors will investigate
further.

We discussed how BBW devices should be addressed. Should they use the FC
address space or something else. Steve suggested that we should view the
B_Port-WAN-B_Port link at a lower level layer than other FC-SW2 operation.
This allows the equivalent of an ISL to be mapped to a lower level link and
allows any BBW addressing to be separated from the FC addressing architecture.
We determined that the address assignement, what ever it is,
should be as dynamic as possible.
All the switch vendors took an action to investigate further.

Given the amount of remaining work, it will be difficult to meet the two week
rule prior to the February meeting to assure a vote for forwardingto
letter ballot. Therefore we determined that April is a more realistic date 
for completion and letter ballot vote. Also, at this point we are adopting a
formal process for FC-BB comments to assure nothing falls through the cracks.
I will post the template in text format on T11 WEB site.

In addition, the error recovery action item that was carried over is closed.

Action Items

1) Provide state machine for flow control - (Brett Ketchum).
2) Investigate adding a SNAP header. (Murali).
3) Addressing of BBWs (ALL).
4) Investigate flow control as it relates to class F frames.
5) FS_TOV implications (ALL).
6) Provide template to formalize FC-BB comment process (Steve).

Next Meeting

Wednesday 2/16/00 at 11AM - 3PM. Huntington Beach, Ca.