[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: DRAFT of the architecture presentation for the WG session this week

Thanks Eric

-  The ? was shorthand for "how do you..." i.e.I was intending to
   pose this as a question!

- the distinction about the characteristics of the identity token space is
  useful, and I've added an additional slide to call this out, as there
  are certainly some implications about structured and unstructured
   spaces in terms of lookup and referral that need to be called out, I agree.

regards,

   Geoff
At 03:45 PM 2/03/2004, Erik Nordmark wrote:
> I may (will) change the presentation between now depending on the response
> from this list,, but I would obviously prefer a little shaking out in the
> wg on the
> more generic topic of "is this architectural taxonomy on the right track" and
> "have I missed an architectural taxonomy element?"

Overall I think this is on the right track.

While in the fullness of time we want session persistence I think we need
to look carefully at what is needed to introduce multihoming when the
peers are not upgraded to use something new. I don't know if
that is what the question mark after "session persistence" on page 11.

Drilling down into the "defining a new protocol element" set there
seems to be two classes:
- those that define a new namespace (HIP, SIM, ...)
- those that use an existing name space such as FQDNs (MAST, NOID, CB64, ...)

With that split in mind it seems odd to clumb together the proposals on page
13; SIM is much closer to HIP (with the biggest difference being that the
payload packets are not encrypted by SIM whereas HIP encrypts everything.)

So understanding which proposals provide a new ID name space to applications
would be useful.

Erik