[183102] in North American Network Operators' Group
Re: Is it possible to roughly estimate network traffic distribution
daemon@ATHENA.MIT.EDU (Martin T)
Fri Aug 14 03:54:43 2015
X-Original-To: nanog@nanog.org
In-Reply-To: <CAPkb-7BrhfiPrO3AE4-dCmhsp1fgF3rZ9x0yxEid02J28XS-nw@mail.gmail.com>
Date: Fri, 14 Aug 2015 10:54:40 +0300
From: Martin T <m4rtntns@gmail.com>
To: nanog@nanog.org
Errors-To: nanog-bounces@nanog.org
Thanks for confirming this! One last question- am I correct that those
graphs referred in my initial e-mail indicate announced prefixes? Only
way to have some insight about received prefixes for particular ASN is
to check the RIR database aut-num object and hope that this is
up-to-date and all the routing policies are describe there in detail?
Again, RIPE Atlas or the NLNOG RING or looking-glass could also help a
little.
thanks,
Martin
On 8/14/15, Baldur Norddahl <baldur.norddahl@gmail.com> wrote:
> You may be able to view what routes I announce but you still have no idea
> what my route policy is like. I might prefer one upstream over another due
> to pricing, latency, capacity or any other unknown reason. And that is
> never published.
>
> If you can not know my egress, you will not know my ingress either as that
> would be someone else egress and you can not know their egress....
>
> You could use RIPE Atlas or the NLNOG RING to do traceroutes. That would
> give you an idea of how traffic actually flows.
>
> Knowing the routes tells you nothing about how much traffic will be
> exchanged. How do you know which ASN has a deal with a big CDN or which
> ASNs are content heavy vs eyeball heavy? Only the source or destination ASN
> can know for sure how much traffic is exchanged.
>
> Regards,
>
> Baldur
>
