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

RE: Jumbo Frames in 10GbE?

Title: RE: Jumbo Frames in 10GbE?

There are a number of ways to standardize Jumbo Frames.  Some much easier and with a shorter timeline than others.  Here's a list of some possible ways:

1) Form a consortium that can take a current solution and create a standard.  This would be similar to what was done for RMII.  The main advantage is that it is collection of companies (rather than individuals, like the IEEE) that are willing to back a proposal and document what it takes to be compliant.  This has a much shorter timeline due to the fact that it doesn't have the overhead of objectives, PAR, 5 criteria, ballots, etc.

2) Create an IEEE standard that changes the maximun FrameSize.  This is similar to what 802.3ac did.  This has the overhead, but it becomes an internationally recognized standard.  The timeline for this development can be relatively short like 802.3ac was.

3) Try to incorporate the change into a standard that is about to be developed, such as the 10GbE standard.  This would have the advantage of being an internationally recognized standard, but given that the current objectives are against this, it would be an uphill battle to change the objectives.  The other problem is that this would greatly increase the scope of work for the 10GbE standard and may adversely affect the timeline of the 10GbE development and the max. FrameSize development.

Although there may be a high demand for Jumbo Frames or larger packet sizes, to try to tie that work into the 10GbE standard development would only serve to create a larger and more complex body of work.  One of the best principles for any standard development that I've seen to date is the KISS principle.  If the HSSG tries to bite off more than it can chew, it will bog down the development of 10GbE.

Given that Jumbo Frames are already being used in the market by a number of companies, the consortium (#1 solution) would probably be the fastest way to generate a standard.  If there is a real need to have it as an IEEE standard, then creating its own standard (#2 solution) would be the best way to go.  The development would be similar to 802.3ac and could probably be done in about the same time span.  Trying to roll it in to the 10GbE is going to be the hardest to do, because there is already an objective that the FrameSize will remain the same and there seems to be a general feeling that this could impact the 10GbE development.

Following the KISS principle, solutions #1 and #2 would be the easiest path.