On Tue, Jun 25, 2013 at 12:27 PM, Templin, Fred L <fred.l.temp...@boeing.com > wrote:
> Hi, > > I just wanted to put out one thought. When talking about fragmentation > and reassembly, never say "always" and never say "never". It is true > that some routers are not well positioned to perform steady-state > fragmentation and reassembly but that does not mean that it should > "never" be used for any purposes. > > "Fragmentation Considered Harmful" was a seminal work, but it fell into > the trap of making black-and-white statements about fragmentation and > reassembly that do not "always" hold in *all* contexts. IP provides: > 1) routing, 2) addressing, 3) a best-effort delivery model, and 4) a > fragmentation and reassembly capability. Omitting 4) would leave us > with something other than IP. > +1 > > Thanks - Fred > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > ipv6@ietf.org > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- >
-------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
