Re: [MEXT] Any Comments fordraft-jeyatharan-mext-flow-tftemp-reference-00

Hi Frank,

Please see below.

BR,
Mohana 

> -----Original Message-----
> From: Frank Xia [mailto:xiayangsong@huawei.com] 
> Sent: Thursday, April 01, 2010 10:48 PM
> To: Mohana Jeyatharan; 'mext'
> Subject: RE: [MEXT] Any Comments 
> fordraft-jeyatharan-mext-flow-tftemp-reference-00
> 
> Hi Mohana
> 
> Please check my replies again..
> 
> BR
> Frank
> 
> -----Original Message-----
> From: Mohana Jeyatharan [mailto:Mohana.Jeyatharan@sg.panasonic.com]
> Sent: Wednesday, March 31, 2010 7:27 PM
> To: Frank Xia; mext
> Subject: RE: [MEXT] Any Comments
> fordraft-jeyatharan-mext-flow-tftemp-reference-00
> 
> Hi Frank,
> 
> Thank you for some quick comments.  
> 
> Please see some replies below.
> 
> BR,
> Mohana 
> 
> > -----Original Message-----
> > From: Frank Xia [mailto:xiayangsong@huawei.com]
> > Sent: Thursday, April 01, 2010 6:19 AM
> > To: Mohana Jeyatharan; 'mext'
> > Subject: RE: [MEXT] Any Comments
> > fordraft-jeyatharan-mext-flow-tftemp-reference-00
> > 
> > Hi Mohana
> > 
> > I had a quick glance.
> > It is interesting. 
> > 
> > Some clarification questions are below:
> > 1)is TFT(Traffic Flow Template)implementation
> >   mandatory for 3GPP mobile node?
> >   I think so from your document.
> 
> If 3GPP network entitiy decides per flow QoS treatment needed 
> for a given prefix/HoA, then dedicated bearers will be 
> created by network and such TFTs will be assigned.  In such a 
> case, the 3GPP MN needs to understand the TFT when sending 
> data.  Thus I would say it is mandatory for the MN. However, 
> it is not mandatory to create dedicated bearers. If there are 
> no dedicated bearers and TFT, the MN will not have TFTs stored in it. 
> Frank=>I am sure about your answer, yes or not. 
> " If 3GPP network entitiy decides per flow QoS treatment 
> needed for a given prefix/HoA ..".  I assume per flow 
> treatment is a prerequisite for 3GPP when we talk about flow 
> binding topic. So, I think your answer is yes. 

Yes, the answer is yes. TFT feature for MN is mandatory.

> > 2)is it has the same descriptive capability as
> >   draft-ietf-mext-binary-ts-04.txt?
> >   I think so.
> The TFT reference traffic selector suboption is different 
> from the binary TS traffic selector sub option. In TFT Ref 
> sub option, the data field only carries the bearer ID to 
> describe the packet filters.
> Multiple packet filters are collapsed to a single bearer ID. 
> Moreover, when there is change in packet filters tied to TFT, 
> a simple refresh flowbinding would do as described in the 
> TFT-ref draft.  Thus there are some subtle differences 
> between the two.
> 
> 
> However, in highlevel both are describing flows when setting 
> filter rules.
> But, what we are highlighting here is that, when TFTs are 
> present at the 3g bearer filtering table, and the MN wants 
> such flows to be sent via 3G path, a simple reference to the 
> TFT is an easier way to set flow binding rules. 
> Frank=>Did you contribute this to 3GPP?  IMHO, TFT is a 
> better way to incorporate into current 3GPP system.

The TFT reference as a new traffic selector sub-option format needs to
be specified in IETF.  Following that it can be used in 3GPP. Yes, using
a TFT to describe the flows in flow binding operation is useful in 3GPP
scenarios as highlighted in the draft. 

> > 
> > BR
> > Frank
> > 
> > -----Original Message-----
> > From: mext-bounces@ietf.org [mailto:mext-bounces@ietf.org] 
> On Behalf 
> > Of Mohana Jeyatharan
> > Sent: Monday, March 29, 2010 7:53 PM
> > To: mext
> > Subject: [MEXT] Any Comments for
> > draft-jeyatharan-mext-flow-tftemp-reference-00
> > 
> > Hi all,
> > 
> > During last IETF-77 MEXT session we could not present the ID 
> > draft-jeyatharan-mext-flow-tftemp-reference-00 and get the direct 
> > feedback from the WG.
> > 
> > Would really appreciate if we can get some comments on this 
> draft and 
> > general opinion about this draft from the WG.
> > 
> > The mechanism highlighted in this draft is a simple 
> mechanism to set 
> > flow binding rules for flows that have descriptors present 
> at the HA 
> > and MN, using a reference to the TFT as a flow descriptor.
> > 
> > Thanks.
> > 
> > BR,
> > Mohana
> > 
> > _______________________________________________
> > MEXT mailing list
> > MEXT@ietf.org
> > https://www.ietf.org/mailman/listinfo/mext
> > 
> > 
> 
> 