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

Re: review of draft-ietf-shim6-failure-detection-03.txt




El 26/04/2006, a las 11:09, <john.loughney@nokia.com> escribió:

Marcelo,

I think we should drop the '... may be used in other contexts.' as
this seems outside of the scope of SHIM6. I don't think that SHIM6
should work on a general purpose failure detection & path exploration

protocol.

I think that there are several scenarios where there seems to be the
need for having a protocol for detecting failures and exploring
alternative paths. For instance, in mobile IP, there is no way to
determine if the current CoA is working and explore alternative coas if

a failure is detected. Similar thing happens in HIP for instance.

So, my take on this is that perhaps it would be good if the failure
detection and alternative path exploration protocols defined by shim6
could be reused in such scenarios (of course if this does not implies
giving up functionality, but i don't think this is the case)

I mean, if we can define a single protocol that can be used in multiple

scenarios that need the same functionality i think it makes sense....

It makes sense, but the devil's in the details.  Many (most?) protocols
running over IP or UDP have some failure detection mechanism - usually
by using of keep-alives.

well, MIP does not have one at the moment and imho this is something needed in there
HIP is in a similar situation AFAIK

  I'm not sure if SHIM6 is the place to do this
general purpose failure detection & path exploration protocol; perhaps
the chairs can comment.

Anyhow, I imagine that SHIM6, MobIKE, HIP, etc. might have slightly
different
needs or requirements for this kind of functionality and how do we
ensure
that we don't over-engineer a solution but make sure it covers the needs
of protocols outside of SHIM6.

i fully agree with this, i would do this only if it is for free (or very reduced cost) i wouldn't want to spend additional effort to accomodate the needs of n different protocols here.

but if we do the faildet protocol and we aimed to be an as general as possible protocol, then maybe the result can be reused somewhere else.

i think we are quite there already, i mean the faildet draft is quite general imho and could easily fit the needs of MIP and HIP AFAICT

Regards, marcelo



John