Re: [MEXT] Important considerations for directionality of flow bindings

[Hesham Soliman <hesham at elevatemobile.com>]

>> => Yes we're discussing the feasibility of such requirement.
>> I.e., what is
>> the implication of such requirement. If you think it's not
>> difficult to
>> solve then I would encourage you to provide arguments for why it's not
>> difficult to solve or write a draft for example.
>> 
> 
> After the requirement is accepted, I'd like to do those works if needed.

=> Part of accepting requirements is to understand what makes sense. This is
a normal discussion for a requirements process. It's not like we're placing
requirements on something new that we've never worked with before, this is a
well-established technology and is well understood.

> 
>>> 
>>> However, the requirement may be useful for flow handover, e.g.
>>> MN/HA has binding two flows to two BID (CoA) respectively.
>>> Based on policies or link overload indication, etc, HA may initiate
>>> flow binding to handoff one of MN's flows from one CoA to another.
>> 
>> => How does the HA know this information? Any pointers to
>> specs or current
>> deployments would be useful.
>> 
> 
> How HA knows this information is out of scope of MIP.

=> Huh? If you want the HA to do something in MIP you need to show an
implementer how to do it.

> For your information, the polcies information could be PCC rules
> downloaded from PCRF which is defined in 3GPP TS23.203
> and TS23.402, etc.

=> I've already said many times that this is not about policies.

Hesham