Re: OT: ip sla packet loss w/o responder ?

From: ALL From_NJ <all.from.nj_at_gmail.com>
Date: Fri, 30 Oct 2009 22:02:37 -0400

Hey Carlos,

Do I understand that you are loosing packets in the cloud? Can we assume
you do not have any visibility?

If your traffic traverses a private network, then the provider should have
some reporting to help you see where the packet loss is. If you are going
over a best-effort public network then good luck ...

So your idea of adding probes via IP SLA might give you some insight as to
where traffic problems are occurring. Will this help you to know?

OER is supposed to address this by having it's own probes as you mention.

For outbound traffic between sites, does it make sense to configure some
shaping and slow that traffic down some? If the far side is getting over
run, then this is not good either.

Sorry if this response is not all that helpful, however, do please respond
back with some additional info and thoughts. Kindest regards,

Andrew

On Fri, Oct 30, 2009 at 9:40 AM, Carlos G Mendioroz <tron_at_huapi.ba.ar>wrote:

> JR,
> I'm trying to monitor packet loss in a real network.
> The whole picture is a medium site with asimetrical connectivity
> to a couple of providers via 4 links, one of them 8:1 bigger
> than the next (26M/4M).
>
> It is congested INBOUND and we are experiencing outbound issues in
> services, I suspect related to inbound packet loss. So I want to track
> loss to be able to put some facts in my theory.
>
> I've been pointed to OER, which in fact does track packet loss w/o
> peer cooperation (as needed AFAIK by ip sla). The problem is that
> I have not been able to find a way to extract that knowledge from it :(
> Border routers talk to controllers using a propietary protocol (DRIP),
> and no MIB is published to expose the data AFAIK.
>
> -Carlos
>
> JR Garcia @ 28/10/2009 14:32 -0300 dixit:
> > you could run a program like ping plotter or something similar on your
> > network.
> > put a pc on your router and do a ping plot to the other side. it will
> graph
> > out latency, drops, ect.
> > not as robust as ip sla, but it works if all your looking for is packet
> loss
> > measurement.
> >
>
> --
> Carlos G Mendioroz <tron_at_huapi.ba.ar> LW7 EQI Argentina
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
>
>
>
>
>
>

-- 
Andrew Lee Lissitz
all.from.nj_at_gmail.com
Blogs and organic groups at http://www.ccie.net
Received on Fri Oct 30 2009 - 22:02:37 ART

This archive was generated by hypermail 2.2.0 : Sun Nov 01 2009 - 07:51:01 ART