Re: [rtcweb] FW: I-D Action: draft-muthu-behave-consent-freshness-01.txt

"Muthu Arul Mozhi Perumal (mperumal)" <mperumal@cisco.com> Wed, 18 July 2012 15:32 UTC

From: "Muthu Arul Mozhi Perumal (mperumal)" <mperumal@cisco.com>
To: Martin Thomson <martin.thomson@gmail.com>
Thread-Topic: [rtcweb] FW: I-D Action: draft-muthu-behave-consent-freshness-01.txt
Thread-Index: AQHNY211pZ5zmVz19UG3CqD7N+7epZcs+oNwgAEHuwCAARjbcA==
Date: Wed, 18 Jul 2012 15:33:14 +0000
Message-ID: <E721D8C6A2E1544DB2DEBC313AF54DE2012D80B3@xmb-rcd-x02.cisco.com>
References: <E721D8C6A2E1544DB2DEBC313AF54DE2012D525F@xmb-rcd-x02.cisco.com> <CABkgnnXOzpnWR6=2DSt7tfpm3VQ9E_cVf4B_8yh6OboPPTOiHQ@mail.gmail.com>
In-Reply-To: <CABkgnnXOzpnWR6=2DSt7tfpm3VQ9E_cVf4B_8yh6OboPPTOiHQ@mail.gmail.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
MIME-Version: 1.0
Cc: "rtcweb@ietf.org" <rtcweb@ietf.org>
Subject: Re: [rtcweb] FW: I-D Action: draft-muthu-behave-consent-freshness-01.txt
Precedence: list

|This looks like a reasonable start.  I'm not sure why there are 
|so many other sections in the document.  It was pretty clear to 
|me that consensus was reached around the use of Binding requests. 
|I'd suggest that you trim the document down to just this section 
|if you intend for it to be adopted.

These sections define a new STUN Consent method. The document will be trimmed if we declare consensus to reuse the STUN Binding method (hopefully after IETF-84, one way or other).

|There needs to be more detail about the nature of the consent/liveness
|test with respect to the observed round trip time and packet loss.

Agree. We will add them in the next revision.

|Based on the discussions at the interim, it was clear that the
|complete 60+ second check time was unnecessary for consent/liveness
|refresh.

STUN requests are retransmitted 7 times by default. With an RTO of 500 ms the transaction would fail after 39.5 ms if there is no response. Our current thinking is this could be limited to 5 retransmissions for consent freshness, so that the transaction fails after 15.5 sec. Does it sound reasonable?

|I can see how you might reach the conclusion that consent is not
|needed for RTCP flows, but I don't see the benefit of creating a
|special case for it.

Yes, the only benefit I can think of is, it reduces some network traffic. 

Muthu

|-----Original Message-----
|From: Martin Thomson [mailto:martin.thomson@gmail.com]
|Sent: Tuesday, July 17, 2012 10:09 PM
|To: Muthu Arul Mozhi Perumal (mperumal)
|Cc: rtcweb@ietf.org
|Subject: Re: [rtcweb] FW: I-D Action: draft-muthu-behave-consent-freshness-01.txt
|
|On 17 July 2012 00:37, Muthu Arul Mozhi Perumal (mperumal)
|<mperumal@cisco.com> wrote:
|> 4. Solution Overview
|>    Discusses the combined consent freshness and session liveness
|>    test (slide "Combined Consent/Liveness Proposal II" from the
|>    RTCWEB interim).
|
|This looks like a reasonable start.  I'm not sure why there are so
|many other sections in the document.  It was pretty clear to me that
|consensus was reached around the use of Binding requests.  I'd suggest
|that you trim the document down to just this section if you intend for
|it to be adopted.
|
|There needs to be more detail about the nature of the consent/liveness
|test with respect to the observed round trip time and packet loss.
|Based on the discussions at the interim, it was clear that the
|complete 60+ second check time was unnecessary for consent/liveness
|refresh.
|
|> 6. Open Items
|>    Lists the current open issues.
|
|I can see how you might reach the conclusion that consent is not
|needed for RTCP flows, but I don't see the benefit of creating a
|special case for it.

[rtcweb] FW: I-D Action: draft-muthu-behave-conse… Muthu Arul Mozhi Perumal (mperumal)
Re: [rtcweb] FW: I-D Action: draft-muthu-behave-c… Martin Thomson
Re: [rtcweb] FW: I-D Action: draft-muthu-behave-c… Muthu Arul Mozhi Perumal (mperumal)
Re: [rtcweb] FW: I-D Action: draft-muthu-behave-c… Martin Thomson