> There are many implementations that don't have the ability to > reassemble a 65,353 byte packet, at least not without allocating > a 64K buffer for this purpose. So, this seems to be a harsh > restriction to place on RFC 2893 implementations. > > If we need to choose one or the other, I'd prefer limiting the > MTU of these tunnels to 1280.
I think the right solution is to strongly recommend path MTU discovery across the tunnel which would have the effect of only using IPv4 fragmentation when the IPv4 path MTU is so small that a 1280 byte IPv6 packet can't be carried without fragmentation. But at least Itojun seems to think that doing path MTU discovery for the tunnel case isn't worth while, and I definitely need to better understand his point. Erik -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
