Re: Reqmts & Init Hdrs drafts

[Mikael Degermark <micke@CS.Arizona.EDU>]

At 12:40 AM 9/8/00 -0700, Bill Manning wrote:

>% >Also, in section 2.1 of the Initialization Headers draft you have
>% >a single bit to discriminate IP versions. There are three active
>% >versions, 4, 5, and 6. Perhaps its short sighted to presume v4 and v6
>% >are the only viable options here.
>% 
>% Hmmm. While what you say is true, our charter only mentions v4 and v6.
>% Moreover, to the best of my knowledge, there is no ST-II expertize in the WG,
>% so we are currently neither required nor able to do anything about v5. 
>
>	While the charter is pretty clear, if you don't allow more bits for 
>	other valid versions of IP, your choice of a single bit will
>	either force a revisit of your work or abandonment of your work.

You might note that the Static part of both kinds of IP header contain 
the 4-bit IP version field, so it is possible to deal with other kinds of IP header, 
the version would be evident from the stored Static header. 

In any case, the Static and Dynamic headers would have to be defined for 
other versions of IP, so it is impossible to avoid a revisit of the work. That said, 
perhaps the 4/6-bit in the header should be removed and instead one should
rely only on the header chain. The bit is really there for efficiency only, when we
did 2507 we discovered that the implementation benefitted from having the version
information up front. 

Mikael Degermark