Re: [MEXT] Scenarios for HA to MN Flow Binding-draft-xia-mext-ha-init-flow-binding

[Frank Xia <[e-mail]>]

Hi Mohana

Thanks for bringing up pramatic scenarios.
Please check my inline comments.

BR
Frank

----- Original Message ----- 
From: "Mohana Jeyatharan" <[e-mail]>
To: "Behcet Sarikaya" <[e-mail]>; <[e-mail]>
Sent: Wednesday, October 07, 2009 7:41 PM
Subject: Re: [MEXT] Scenarios for HA to MN Flow 
Binding-draft-xia-mext-ha-init-flow-binding


> Hi Behcet and all,
>
> From the use cases people have proposed for flow binding work, it is
> acceptable that in most cases, the decision about the flow binding rule
> is made by the MN.
Frank=>I would like to say SOME people.  AFAIK, there are also some
people favor HA initiating flow binding.  You can tell from the verbose
discussion.

>
> But as mentioned in draft-xia-mext-ha-init-flow-binding, the flow
> binding rule can originate from HA(be it mandatory or optional).  Bcos
> in some cases, the network can decide what is best for it.  Since there
> is a valid trust relation ship between HA and MN, the HA can use flow
> binding protocol in a reverse way(as quoted in HA-init ID) to sends its
> rules to MN.
Frank=>Thank for mentioning our draft!

>
> In addition to the scenarios you have quoted,  I am quoting another
> scenario, where the HA can suggest its flow binidng rules and the MN can
> use these rules to create its own rules.
> ****************************************************
> Flow binding rules initiated by HA
> and their categories
>
> [1]Some are Mandatory flow binding rules-case A
> [2]Some are flow binding preferences and UE can decide on the suitable
> flow binding rule based on some preference given by the network/HA-case
> B
> [3]Some flow binding rules are mandatory but to be applied when an event
> occurs-Case C
>
> ***************************************************
> Case A and Case B scenarios are given together
> ***************************************************
> HA changing default flow policies into default flow binding
> ============================================================
>         +-----+
>         | HA  |
>         |     |
>         +-----+
>          /    \
>    3G homelink \
>   (PMIP)/       \
>        /         \ HPLMN
>       /           \
>     +-----+    +-------+
>     | S-GW|    | ePDG  |
>     |     |    |       |
>     +-----+    +-------+
>         \        /
>          \      /
>           \    /
>            +---+
>            |MN |
>            +---+
> The network entity such as HSS through AAA gives default flow policies
> to HA (Voice via 3G and data via other).  The HA changes default
> policies
> into default flow binding rules.
> The HA has PMIP binding for the home link and knows the binding for home
> link and the access type tied to homebinding.
> home binding BID1, foreign binding BID2.
>
> Thus HA initiates default flow binding rule (FID1(voice) BID1.
> FID2(data) BID2.
> MN generates flow binding rules for uplink and downlink traffic based on
> flow binding rules given by HA.   HA given rules can be subset of the
> final MN flow binding rules.
>
> In case A, MN needs to follow these rules strictly.
>
> In case B, if there is no other external factor influencing the flow
> binding rules being set by MN(ANDSF, MNs preferences etc) it will follow
> the HA given flow binding rule.
> Else it will NACK the HA given flow binding rule and give its own flow
> binding rule.
> *******************************************************************

Frank=>We presented this use case last IETF meeting.  Many argued  that this 
is about
default policy not binding rule provisioning, and this can be done through 
higher layer
mechnisms other than layer 3 solution.   Of course, I dont agree.  Flowing 
binding
is an operation above layer 3.   The terminologies demarcation in this 
context is
confusing.


> Comments?
>
> Thanks.
>
> BR,
> Mohana
>
>> -----Original Message-----
>> From: [e-mail] [mailto:[e-mail]] On Behalf
> Of
>> Behcet Sarikaya
>> Sent: Wednesday, September 30, 2009 3:45 AM
>> To: [e-mail]
>> Subject: [MEXT] Scenarios for HA to MN Flow Binding
> -draft-xia-mext-ha-
>> init-flow-binding
>>
>> Hello all,
>>   As presented in IETF 75
>> http://www.ietf.org/proceedings/75/slides/mext-2/mext-2.htm
>>
>> apart from flow binding revocation, we had
>> - inter-interface flow movement and
>> - default Flow Binding Provisioning
>> scenarios.
>>
>> Please post any comments on the above or related to the above
> scenarios.
>>
>> Regards,
>>
>> Behcet
>>
>>
>>
>>
>> _______________________________________________
>> MEXT mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/mext
> _______________________________________________
> MEXT mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/mext