| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Kevin,
I could vote for 1875 or 2048, it depends whether we're bound to backwards compatibility or not. Do we have guidance from 802.3 which could help steer us?
...Dave
David W. Martin
Nortel Networks
dwmartin@ieee.org
+1 613 765-2901 (esn 395)
~~~~~~~~~~~~~~~~~~~~
-----Original Message-----
From: Kevin Daines [mailto:Kevin.Daines@WWP.COM]
Sent: Thursday, September 02, 2004 3:11 PM
To: STDS-802-3-FE@LISTSERV.IEEE.ORG
Subject: [FE] Frame size poll
All,
I'd like to poll the reflector and see what folks are thinking in terms of the new maximum frame size. Before I do, first a bit of background for those that haven't attended recent 802.1, 802.3, or 802.3 Frame Expansion Ad Hoc meetings.
802.1 has two projects that require additional octets. First, 802.1ad Provider Bridging has requested an additional four-octet VLAN tag. 802.1AE MACSec requests some number of header (SecTag) octets, perhaps 4 or 8, and trailer (ICV) octets, perhaps 64-256. In addition, ITU-T is defining MPLS encapsulation requiring a minimum of 30+ octets.
802.3 generally acknowledges the need to increase the frame size, but prefers to only do it one time.
802.1 and ITU-T have asked for additional octets to cover their current needs and also requests knowing what the "upper bound" is. Therein lies one of the problems facing the FESG. Per the specification, the upper bound is 1522 octets. Pat Thaler presented in July some data that suggests most repeater implementations "likely" support 357 additional octets (for a total of 1875) based on a round number of bits being buffered to account for clock tolerances. So, 1875 has emerged as one candidate max frame size. Others, have suggested we make it a power of 2: 2,048 and "call it good."
Now to the poll. Would you prefer:
1) 1875 (which is based on the repeater calculation)
2) 2048 (a power of 2)
3) some other number (please specify)
Thanks. I'll tally and provide a summary when folks have had a chance to weigh in.
FYI: Al Braga from UNH IOL has volunteered to test the gear in their lab and determine some upper bounds. This should provide the SG some good data about existing gear. If anyone else wants to provide data, please let me know.
Kevin Daines
Chair, 802.3 Frame Expansion Study Group