Ethereal-dev: RE: [Ethereal-dev] SCTP stat and Agentx dissector

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: "Terletsky Oleg" <oleg.terletsky@xxxxxxxxxxxx>
Date: Wed, 9 Feb 2005 17:11:23 -0500
I called it end-point, but think more correct is to collect statistics
per associations. 
About retransmissions, I think it's easier to check some counter instead
of analyze TSN sequencing.

Thanks,
Oleg.

-----Original Message-----
From: Michael Tuexen [mailto:Michael.Tuexen@xxxxxxxxxxxxxxxxx] 
Sent: Wednesday, February 09, 2005 4:38 PM
To: Terletsky Oleg
Cc: ethereal-dev@xxxxxxxxxxxx; guy@xxxxxxxxxxxx
Subject: Re: [Ethereal-dev] SCTP stat and Agentx dissector

Hi Oleg,

see my comments in-line.

Best regards
Michael

On Feb 9, 2005, at 22:32 Uhr, Terletsky Oleg wrote:

> Hi Michael,
>
> The reason I added SCTP conversation is because I found it very useful

> for measuring retransmissions on mutihomed association.
> It helps to identify network configuration problems.
> With regards to filters I didn't finish it yet. But if your sctp_stat 
> already doing it I'm not going to do it.
>
> The bottom line is that we needed counters for each chunk sent and 
> received from/to each SCTP endpoint.
Do you need this per association or per end-point?

The analysis in Ethereal can plot TSNs over the time. So you can see
retransmissions...
>
> Thanks,
> Oleg.
>
> -----Original Message-----
> From: Michael Tuexen [mailto:Michael.Tuexen@xxxxxxxxxxxxxxxxx]
> Sent: Tuesday, February 08, 2005 5:07 PM
> To: Terletsky Oleg
> Cc: ethereal-dev@xxxxxxxxxxxx; guy@xxxxxxxxxxxx
> Subject: Re: [Ethereal-dev] SCTP stat and Agentx dissector
>
>
> Hi Oleg,
>
> I have played with the SCTP code you sent and have some comments: 1. 
> How
> much sense does it make to have SCTP conversations?
>     I'm wondering, because for UDP or TCP this corresponds to
>     a TCP connection or a UDP communication relation. But for
>     SCTP it has no sense, since SCTP associations are multihomed.
>     Guy, what do you mean?
> 2. If the user specifies a display filter for an association
>     you provide a statistic for the chunks. Currently on the last
>     chunk within each packet matching the display filter is counted.
>     How do you use this feature? Do you need this counter per
>     association for each direction separately or do you
>     use also other filters? We could integrate such a statistic
>     in the SCTP analysis window which shows information per 
> association.
>
> Best regards
> Michael
>
> On Feb 7, 2005, at 20:06 Uhr, Terletsky Oleg wrote:
>
>> Hi,
>>
>> With regards to SCTP stat. I think it's great that you already have 
>> something. We will start using it as soon as you publish it.
>>
>> It just currently we in the middle of the testing some big project 
>> that includes SCTP measurement stuff and I found it very useful for 
>> our testers instead of struggle with bunch of filters just get all 
>> information in one shot.
>>
>> Guy,
>> Another question is AgentX dissector.
>> I sent the tarball and I'd like a to get some comments on it also.
>>
>> Thanks,
>> Oleg.
>>
>>
>> -----Original Message-----
>> From: Terletsky Oleg
>> Sent: Friday, February 04, 2005 4:58 PM
>> To: 'ethereal-dev@xxxxxxxxxxxx'
>> Cc: 'guy@xxxxxxxxxxxx'
>> Subject: [Ethereal-dev] SCTP stat and Agentx dissector
>>
>>
>> Sorry for the wrong subject in the prev. message.
>>
>>
>> Hi,
>> I'd like contribute another work to ethereal community.
>>
>> 1) I've added some statistics to SCTP.
>>    tap-sctp.tgz contains the following:
>>
>>    gtk/sctp_stat.c - contains SCTP stat. GUI
>>    tap-sctpstat.c  - tethereal variant of the same stat.
>>    gtk/conversation_sctp.c - impl. of SCTP conversation info.
>>
>>    tap-sctp.patch - has a changes for packet-sctp.[hc] and 
>> tap-iousers.c and also for list of CM files
>>    in order to support sctp stat and AgentX plug-in.
>>
>> 2) Plug-in dissector for AgentX protocol.
>>
>> I've tested it on Linux RHAS 3.0 It builds and executes fine.
>>
>> Thanks.
>> Oleg.
>>
>>
>
> ______________________________________________________________________
>   This email message has been scanned by PineApp Mail-Secure and has 
> been found clean.
>
>

______________________________________________________________________
  This email message has been scanned by PineApp Mail-Secure and has
been found clean.