Issues with frame expansion

IEEE group’s concerns with enlarging Ethernet frames, alternative approaches

Last time, I looked at how a committee within the IEEE 802.3 group has started to consider changing the maximum frame size allowed in Ethernet networks, a change request brought up by another group in the IEEE. This time, let’s look at the issues such a change raises.

The number-one question the committee raised was, “Is this absolutely necessary?” The maximum frame size underlies so many other technologies that changing it might have unintended consequences. That’s why the second concern listed by the committee was backwards compatibility.

As alternatives, the group mentioned these approaches:

To continue reading, register here and become an Insider. You'll get free access to premium content from CIO, Computerworld, CSO, InfoWorld, and Network World. See more Insider content or sign in.

Last time, I looked at how a committee within the IEEE 802.3 group has started to consider changing the maximum frame size allowed in Ethernet networks, a change request brought up by another group in the IEEE. This time, let’s look at the issues such a change raises.

The number-one question the committee raised was, “Is this absolutely necessary?” The maximum frame size underlies so many other technologies that changing it might have unintended consequences. That’s why the second concern listed by the committee was backwards compatibility.

As alternatives, the group mentioned these approaches:

“* Exclude .3 support from .1ad & .1AE standards.
* Handle on bootleg basis.
* Require .1 SAR function.
* Discover MTU size capability.
* Reduce client MTU to make room for added headers.”

Also, if the IEEE goes ahead with extending the frame size, it may be desirable to limit use of the bigger frames to just those links that require them, using discovery to find those links.

The group also questioned the impact of a change on quality of service, efficiency, and in particular, jitter.

From presentations on the IEEE Web site, it looks like the committee will next research several areas, such as the frame size limitations of components of today’s equipment - things like elasticity buffers and delimiters. It will also look into the effect of increased overhead on performance, the use of 802.1AB for discovery and the feasibility of reducing the MTU of the installed base of clients.

Yes, this is potentially a big change, but before you go writing your congressman about it, bear in mind that it’s still early days.

Read more about lans & wans in Network World's LANs & WANs section.