Re: [Tsvwg] Re: The requiremntsand proposalsforM2PA/M3UA/SCTPExtension from China Mobile

Ronnie,

That's interesting.  I don't actually do that.  I send the HEARTBEAT-ACK
to the Source IP in the HEARTBEAT, provided that I have a usable route to
that address, I have no retransmissions that were originally sent to
that address outstanding, and I have not received duplicate TSNs when
SACKs were sent to that destination.  Otherwise, I send them to the
current primary.

--brian

[email protected] wrote:                    (Mon, 30 Jul 2007 04:02:25)
> Hi,
> 
> Won't this association have a serious problem with Heartbeats if you
> cannot send a Heartbeat Ack to the address the Heartbeat came from?
> 
> >From RFC 2960bis
> 	3.3.6. Heartbeat Acknowledgement (HEARTBEAT ACK) (5): 
> 
> 	An endpoint should send this chunk to its peer endpoint as a
> response to a 	HEARTBEAT chunk (see Section 8.3). A HEARTBEAT
> ACK is always sent to the source IP 	address of the IP datagram
> containing the HEARTBEAT chunk to which this ack is responding.
> 
> We also need to take into account that the source address of duplicate
> data may NOT be the source address used for the original data if the
> sending end has changed paths when the data wasn't acked.
> 
> If this is a routing problem which has been in place since the
> association was started then the sender should never send data to an
> unconfirmed address.  If the Heartbeat mechanism is broken the address
> will never be confirmed.
> 
> Ronnie
> ---
> Ronnie Sellar			Emerson Network Power, Embedded Computing
> Tel: +44 131 475 7015		Suite 4, 2 Anderson Place
> Fax: +44 131 475 7001		Edinburgh EH6 5NP
> mailto:[email protected]	http://www.emersonembeddedcomputing.com
>  

-- 
Brian F. G. Bidulock
[email protected]
http://www.openss7.org/