On Sun, 14 Mar 2010 14:19:22 +1300 Brian E Carpenter <[email protected]> wrote:
> On 2010-03-14 13:46, Mark Smith wrote: > ... > > These are the sorts of scenarios which have made me wonder whether > > there should be some sort of mechanism to express different off-link > > and on-link MTUs or on-site and off-site MTUs (distinguished by an > > on-site aggregate prefix). > > I hope they're discussing that over in MIF. A different set of settings > (not just MTU) per prefix seems entirely rational; a nice large MTU > for a ULA prefix might make sense. > I've thought you could go a bit further, allowing nodes to have different MTUs on the local link. The current RA announced MTU is made the offlink unicast/multicast MTU, onlink multicast MTU and default MTU, if the following options aren't understood. E.g. for the 6RD scenario, this would be 1280 bytes. Another RA option is created to announce the maximum MTU of the link, being the largest that the link supports e.g. the maximum frame size the link layer can forward. If all switches were configured to support 9K byte frames, this value would be 9K. This would default to the normal or common link layer MTU values e.g. 1500 bytes for ethernet. A ND NA option is created for a node to announce it's local MRU. A sending node that understands these RA/ND NA options would then be able to send unicast packets up to the smallest value of it's local MTU, the maximum link MTU and the neighbor's MRU. E.g. if the local interface only supports 8K MTU/MRU, then in that node's NAs it would announce that 8K MRU, and other nodes that supported 9K MTUs would only send 8K frames to it. Regards, Mark. _______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
