Re: [Pana] PANA-Error-Request

[Yoshihiro Ohba <yohba at tari.toshiba.com>]

On Fri, Oct 27, 2006 at 08:47:06PM +0300, Alper Yegin wrote:
> I see.
> 
> Than my second thought was, why not carry the whole PANA header? Why is the
> bits carried is limited to what you stated?

This is is because I thought only 'R' bit of Flags, Message Type and
Sequence Number fields are sufficit for message identification.
Version and Message Length fields would not be part of message
identification information.  Session Identifier is not needed because
Session Identifier is the same as that is contained in PER.

However, if the usage of the Failed-Message AVP is more than message
identification, it would make sense to carry the whole PANA header.

Yoshihiro Ohba



> 
> Alper
> 
> > -----Original Message-----
> > From: Yoshihiro Ohba [mailto:yohba at tari.toshiba.com]
> > Sent: Friday, October 27, 2006 5:03 PM
> > To: Alper Yegin
> > Cc: 'Yoshihiro Ohba'; pana at ietf.org
> > Subject: Re: [Pana] PANA-Error-Request
> > 
> > Alper,
> > 
> > Since sequence numbers used for requests generated by PaC and
> > generated by PAA are independent, it is possible that the sequence
> > numbers contained in two consecutive messages sent by the same entity
> > can be the same.
> > 
> > For example, if the PaC sends a PAN (without EAP-Payload) in response
> > to a PAR generated by the PAA and then sends a PAR (with EAP-Payload),
> > then the PAN contains the sequence number assigned by the PAA and
> > contained in a PAA-generated PAR while the PaC-generated PAR contains
> > the sequence number assigned by the PaC, the two sequence numbers can
> > happen to be the same (perhaps with a very low probability).  As a
> > result, when a PER for one of those messages are generated by the PAA,
> > the PaC will not able to determine which message caused an error.
> > 
> > So I think Sequence Number is not sufficient to determine the message
> > that caused the error.
> > 
> > Yoshihiro Ohba
> > 
> > 
> > On Fri, Oct 27, 2006 at 01:44:05PM +0300, Alper Yegin wrote:
> > > >    The Algorithm AVP (AVP Code X) is of type Unsigned64 containing
> > > >    Flags, Message Type and Sequence Number fields of the header of the
> > > >    message that caused the error.  The AVP Data field is encoded as
> > > >    follows:
> > > >
> > > >     0                   1                   2                   3
> > > >     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
> > > >    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> > > >    |            Flags              |         Message Type          |
> > > >    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> > > >    |                      Sequence Number                          |
> > > >    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> > > > "
> > >
> > > Actually, why not simply carry the sequence number only. This is
> > sufficient
> > > to identify a unique message within a PANA session.
> > >
> > > Alper
> > >
> > >
> > >
> > >
> > > >
> > > > Yoshihiro Ohba
> > > >
> > > > _______________________________________________
> > > > Pana mailing list
> > > > Pana at ietf.org
> > > > https://www1.ietf.org/mailman/listinfo/pana
> > >
> > >
> 
> 
> _______________________________________________
> Pana mailing list
> Pana at ietf.org
> https://www1.ietf.org/mailman/listinfo/pana
> 

_______________________________________________
Pana mailing list
Pana at ietf.org
https://www1.ietf.org/mailman/listinfo/pana