Interworking M2PA/SCTP over NxT1 MLPPP

Does anyone have experience interworking M2PA/SCTP over NxT1 MLPPP?
Any recommendations to ensure smooth operation?
Anything to watch out for?

_______________________________________________
Sigtran mailing list
Sigtran <at> ietf.org
https://www1.ietf.org/mailman/listinfo/sigtran

SG M2UA handling of ASP switchover

Hi,
In case of an ASP switchover scenario in an M2UA redundant-ASP setup, I
am facing the below issue -

-1- SG is configured for redundant ASPs (ASP1 and ASP2), ASP1 is the
Active ASP.

-2- ASP1 goes Down, ASP2 takes over; it sends a Data Retrieval request
to SG to retrieve the buffers.

-3- Should MTP3 (corresponding to ASP1) on MGC send a Release Request
for the link just before going down? I assume it should not, but in case
it does send the Release Request for a link just before going down, is
the SG expected to buffer all the data to be transmitted till the new
ASP takes over and sends the Retrieval request?

-4- In the above case, can the SG start the recovery timer [T(r)] on
getting the Release Request for the link, to queue/buffer data for later
retrieval by MTP3 of ASP2?

-5- If T(r) is used as mentioned above, would it be a violation of the
RFC (I did not find a similar scenario in RFC 3331).

Any help would be appreciated!
Regards,
Sunita

Re: SG M2UA handling of ASP switchover

Sunita,

Sunita Bhagwat wrote:                                                               (Wed, 06 Apr 2005 19:04:08)
> Hi,
> In case of an ASP switchover scenario in an M2UA redundant-ASP setup, I
> am facing the below issue -
> 
> -1- SG is configured for redundant ASPs (ASP1 and ASP2), ASP1 is the
> Active ASP.
> 
> -2- ASP1 goes Down, ASP2 takes over; it sends a Data Retrieval request
> to SG to retrieve the buffers.

I suppose you mean that the SG sends alternate ASP active and starts
routing traffic to ASP2.

Why does ASP2 send a Data Retrieval request?

> 
> -3- Should MTP3 (corresponding to ASP1) on MGC send a Release Request
> for the link just before going down? I assume it should not, but in case
> it does send the Release Request for a link just before going down, is
> the SG expected to buffer all the data to be transmitted till the new
> ASP takes over and sends the Retrieval request?

No. Release Request should not be sent outside of normal Q.703/704
procedures.

> 
> -4- In the above case, can the SG start the recovery timer [T(r)] on

Re: SG M2UA handling of ASP switchover

Hi Brian,
Thanks for the quick response. My reply inline..

On Wed, 2005-04-06 at 19:35, Brian F. G. Bidulock wrote:
> Sunita,
> 
> Sunita Bhagwat wrote:                                                               (Wed, 06 Apr 2005 19:04:08)
> > Hi,
> > In case of an ASP switchover scenario in an M2UA redundant-ASP setup, I
> > am facing the below issue -
> > 
> > -1- SG is configured for redundant ASPs (ASP1 and ASP2), ASP1 is the
> > Active ASP.
> > 
> > -2- ASP1 goes Down, ASP2 takes over; it sends a Data Retrieval request
> > to SG to retrieve the buffers.
> 
> I suppose you mean that the SG sends alternate ASP active and starts
> routing traffic to ASP2.
> 
No, AS is configured in an override mode, so SG would receive the
ASP-Active msg from ASP2.

> Why does ASP2 send a Data Retrieval request?
> 
I presume since both the ASPs would correspond to different MTP3 layers
at the MGC, this is a case of MTP3 changeover? The dying MTP3 sends a
Release Request, and the new one is sending the Data Retrieval request.

> > 

sctp heartbeat timer

Hi,

I have one basic question regarding the SCTP heartbeat timer value.
For the regular SG-ASP deployment using M3UA, is it realistic if we
set the default heartbeat timer as 30 seconds suggested in RFC? I mean
if it takes 30 seconds to for SCTP user to realize the path failure,
can M3UA user such as ISUP tolerate such long delay?

Please clarify my confusion,

thanks,
Henry

RE: sctp heartbeat timer

You can set this value lower, depending on your network characteristics. The
"recommended" values in IETF RFCs foresee deploment in big-I Internet and
try to protect the network for general case.

   Tolga

> -----Original Message-----
> From: sigtran-bounces <at> ietf.org [mailto:sigtran-bounces <at> ietf.org]On
> Behalf Of Henry Wong
> Sent: Wednesday, April 06, 2005 12:12 PM
> To: sigtran <at> ietf.org
> Subject: [Sigtran] sctp heartbeat timer
>
>
> Hi,
>
> I have one basic question regarding the SCTP heartbeat timer value.
> For the regular SG-ASP deployment using M3UA, is it realistic if we
> set the default heartbeat timer as 30 seconds suggested in RFC? I mean
> if it takes 30 seconds to for SCTP user to realize the path failure,
> can M3UA user such as ISUP tolerate such long delay?
>
> Please clarify my confusion,
>
> thanks,
> Henry
>
> _______________________________________________
> Sigtran mailing list
> Sigtran <at> ietf.org

Re: SG M2UA handling of ASP switchover

Sunita,

It is not an MTP changeover.  By analogy to Q.704 it is a short term
processor outage.

--brian

Sunita Bhagwat wrote:                                                               (Wed, 06 Apr 2005 20:12:57)
> Hi Brian,
> Thanks for the quick response. My reply inline..
> 
> On Wed, 2005-04-06 at 19:35, Brian F. G. Bidulock wrote:
> > Sunita,
> > 
> > Sunita Bhagwat wrote:                                                               (Wed, 06 Apr 2005 19:04:08)
> > > Hi,
> > > In case of an ASP switchover scenario in an M2UA redundant-ASP setup, I
> > > am facing the below issue -
> > > 
> > > -1- SG is configured for redundant ASPs (ASP1 and ASP2), ASP1 is the
> > > Active ASP.
> > > 
> > > -2- ASP1 goes Down, ASP2 takes over; it sends a Data Retrieval request
> > > to SG to retrieve the buffers.
> > 
> > I suppose you mean that the SG sends alternate ASP active and starts
> > routing traffic to ASP2.
> > 
> No, AS is configured in an override mode, so SG would receive the
> ASP-Active msg from ASP2.

Re: sctp heartbeat timer

Henry,

Setting this value to 30 or 60 seconds does not mean that packets will
not be retransmitted for 30 or 60 seconds, under normal operating
conditions, the Duplicate and Fast Retransmission algorithms can trigger
a path reselection or retransmission long before the 30 or 60 second
timeout.

Besides 30 seconds will not be the path failure timeout, it will be
30 x (2^n - 1) seconds, where `n' is the Path.Max.Retrans.  With a
recommeded value of 5, that will be about 8 minutes.

--brian

Henry Wong wrote:                            (Wed, 06 Apr 2005 12:12:07)
> Hi,
> 
> I have one basic question regarding the SCTP heartbeat timer value.
> For the regular SG-ASP deployment using M3UA, is it realistic if we
> set the default heartbeat timer as 30 seconds suggested in RFC? I mean
> if it takes 30 seconds to for SCTP user to realize the path failure,
> can M3UA user such as ISUP tolerate such long delay?
> 
> Please clarify my confusion,
> 
> thanks,
> Henry
> 
> _______________________________________________
> Sigtran mailing list

Re: sctp heartbeat timer

Henry Wong,您好！

	I think the Heartbeat function is same as the SLTM and SLTA function in MTP3
the SCTP not only use  Heartbeat to determine the reachable to destination ,but also 
use the ack function to to determine the reachable to destination .
so when the   Heartbeat function fail, the SCTP association will be terminated,
if all the message no response from remote side, the Association will be terminated alseo.

======= 2005-04-07 00:12:07 您在来信中写道：=======

>Hi,
>
>I have one basic question regarding the SCTP heartbeat timer value.
>For the regular SG-ASP deployment using M3UA, is it realistic if we
>set the default heartbeat timer as 30 seconds suggested in RFC? I mean
>if it takes 30 seconds to for SCTP user to realize the path failure,
>can M3UA user such as ISUP tolerate such long delay?
>
>Please clarify my confusion,
>
>thanks,
>Henry
>
>_______________________________________________
>Sigtran mailing list
>Sigtran <at> ietf.org
>https://www1.ietf.org/mailman/listinfo/sigtran


= = = = = = = = = = = = = = = = = = = =
			

Query on linkset statistics.

Hi !
Let us see the following linkset statistics structure.

typedef struct _Mtp3LinkSetStats  /* MTP Level 3 link set 
statistics                 */
{
  U32  strtLnkSetFail;   /* Start of linkset failure           */
  U32  stopLnkSetFail;   /* Stop of linkset failure            */
  U32  durLnkSetUnav;    /* Duration of linkset unavailability */

} MTP3LinkSetStats;

what do strtLnkSetFail and stopLnkSetFai mean.
I read it is number of start and stop of linkset failures. Please clarfy it.

Thanks in Advance!
Regards,

Arvind