There is actually a small amount of value from this, even absent a way to distinguish unknown extension headers and unknown upper layer headers. Specifically, the implementation of a parser for (known) extension headers can be simplified due to the uniform layout of at least the start of the header. In software, you could use a single subroutine to start the parsing of any extension header. Hardware implementations would have a similar advantage. While at present most extension headers can use the same subroutine, that is less certain for the future unless a standard format is specified.
At present, when people are thinking about parsing extension headers, I always tell them to be careful not to treat parsing the Fragment header the same way as other defined extension headers, since a non-zero value in the second octet of that needs to be treated differently than the other cases (i.e., it does not indicate a longer length). It would nice to no longer have to give this advice. -- Jim -- Jim Hoagland, Ph.D., CISSP Principal Security Researcher Advanced Threat Research Symantec Security Response www.symantec.com -------------------------- Office: (650) 527-0946 On 3/20/08 1:30 PM, "Vishwas Manral" <[EMAIL PROTECTED]> wrote: > Hi Markku, > > I totally agree with you on this. > > As there is no way to distinguish between an IPv6 extension header and > an upper layer header (TCP/ UDP) there is no way to find this out. > Also the upper layer header is not be IPv6 specific, but common for > IPv4 and IPv6. > > Thanks, > Vishwas -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------