See below

> Hi Mirja,
>
>>>>
>>>> Having said this, I'd like to first see a complete section (3.2) on
>>>> MPTCP before we start to writing something on this in section 4.
>>>> However, I'm sure you could also help to provide some text in section
>>>> 3.2...? That would be great!
>>>
>>> Yes, that was the plan. I think the API for both SCTP and MPTCP are
>>> relevant in highlighting the underlying features of the protocol, even
>>> though these APIs are not what needs to be described.
>>
>>Yep. APIs should not be discussed in section 4. However, for the protocol
>>description in section 3, if you look at the other descriptions, there is
> also a
>>subsection on the (higher layer) interface. As you say there is often a
> strong
>>dependency therefore I think there is a purpose to describe the interface
> as
>>well (in section 3) to have a ground truth for discussion.
>>
>>>
>>> The synthesis in section 4 should come at a later stage, once 3.2 (and
>>> perhaps a similar discussion in SCTP's section), have been written up.
>>
>>Yes!
>>
> [Karen ] Do you here refer to SCTP MH or CMT SCTP ?
>
> MPTCP include concurrency aspects which, for SCTP, only significantly
> arise with CMT SCTP.
>
> It has previously been stated that CMT SCTP would not be in scope of taps.
> Is that still the assumption ?
>
> BR, Karen
>
>>Mirja
>>

I think SCTP multihoming and path failover should be there.

I personally think that formal discussion of SCTP/ CMT falls outside the
scope of TAPS, until the IETF decides to adopt this work (that's for TSVWG
sometime). However, I guess if TAPS wanted to refer to a paper on the
possibility of such extensions to SCTP in the discussion section that
could easily be considered.

Gorry


_______________________________________________
Taps mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/taps

Reply via email to