SCTP Discussions

Text archives Help


Re: [SCTP-Discussion] About detection of path failure


Chronological Thread 
  • From: Michael Tüxen <Michael.Tuexen AT micmac.franken.de>
  • To: hektor AT etri.re.kr
  • Cc: discussion AT sctp.de
  • Subject: Re: [SCTP-Discussion] About detection of path failure
  • Date: Sat May 3 22:24:34 2003
  • List-id: SCTP Discussions <discussion.sctp.de>

Hi,

see my comments below.

Best regards
Michael

On Wednesday, Apr 30, 2003, at 06:44 Europe/Berlin, hektor AT etri.re.kr wrote:

G'day.

About Path failure detection

According to SCTP specification (RFC2960), paths' status maintained by
HeartBeat mechanism.
But In case of the primary path, which means currently used to exchange data, HeartBeats
are not placed without explicit Heartbeat reqeust.
the primary path's status is just checked by real data exchanges and related retransmission
mechanism.
Is that right?

No. RFC 2960 does not make any difference between the primary path and the others.
It only makes a difference between idle paths and paths which are not idle. If
a path (primary or not) becomes idle, it is supervised by the HEARTBEAT procedures.
If yes, then If after association established, and there's no data exchange,
How to detect failure of the primary path?
Is the SCTP specification dealing with this? Is that just the matter of implementation?

RFC 2960 is as described above and the sctplib should do this. Have you found that it
behaves differently?

Thanks In Advance.





  • Re: [SCTP-Discussion] About detection of path failure, Michael Tüxen, 05/03/2003

Archive powered by MHonArc 2.6.19+.

Top of page