Veera Tubati (vtubati) writes: > > We're talking about a single MTU-sized packet transmitted and one > > received on an Ethernet link. > > Correct, but it could be in the order of 20k or more sessions trying to > come up at once after the outage of a BRAS box. For the broken clients > asking for 1500, it is more overhead to timeout and retry something > else. Whatever can be avoided reasonably is a bonus. So ... we're worried about a BRAS that is unable to handle the load of one extra packet for those 20K clients but that will perform "acceptably" when those same 20K clients are all actively using their links? I don't quite understand, but it sounds to me like a design issue for the BRAS and not a protocol issue. If it wants, that machine could sequence the link bring-up so that it spreads out the load, or it could just use a more capable hardware platform. > May be there would be cases where ISP is sure that their underlying > network and clients support 1500 MRU indeed... I agree that consenting peers may do as they like, but it's not an optimization that makes sense to me. -- James Carlson, KISS Network <james.d.carlson@xxxxxxx> Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084 MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677 _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf