Earlier, Manav wrote: > Assume you are the end host that wants to prioritize certain packets > or wants to implement Access control lists (ACLs). In the absence > of this extension a router cannot apply ACLs as it will never know how > to parse the packet in case it comes across an unknown extension header.
Banning new extension headers and instead requiring use of existing *already supported by multiple routers* IPv6 headers is a much better, much more complete solution to the situation postulated above. It is practical to use the existing IPv6 headers. The existing headers are well designed and widely supported in deployed nodes today. I've specified enhancements using the existing headers twice so far. RFC-5570 is one example, while draft-rja-ilnp-nonce is another example. By contrast, specifying this new extension header both encourages new extensions (which is undesirable, as Joel H keeps observing) and more importantly *breaks* existing IPv6 deployments that use commercial routers that already can parse past the existing IPv6 extension headers to view transport-layer information (e.g. protocol and port numbers). Yours, Ran -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
