[c-nsp] Billing per Mb (Large ISPs)

[Blake Dunlap]

The reason is because as an ISP, you generally have to build out for
peak, not average use, unless you're [maligned eyeball provider] and
just don't care.


-Blake

On Sun, May 25, 2014 at 2:22 AM, Saku Ytti <saku at ytti.fi> wrote:
> On (2014-05-24 16:44 -0500), Blake Dunlap wrote:
>
>> For those that don't realize the underlying math: Per mb billing is
>> based on average throughput for the period, 95th is billing based on
>> lopped ceiling peak.
>
> Aye. per MB is much more fair, two 95 percentile customers could be using
> wildly different amount of network capacity yet pay exactly the same. I wonder
> what is the history behind 95 percentile, why did we make it de-facto for IP
> transit?
>
> I'm also interested if Australia has similar content/eyeball disputes as EU
> and US for those companies who do per byte billing?
>
> Intuitively it would seem that if I'm being paid for bytes I'm delivering, I'd
> want as many bytes from content networks as I can possibly get. As opposed to
> if I'm not being paid, I'd like to deliver as little as possible, to avoid
> eroding margins.
>
> This is what we're seeing in EU and US, eyeball networks regularly just stop
> delivering content bytes until content shops pays to them. Which creates
> inefficient indirect billing as netflix et.al. then have to increase end-user
> pricing to cover these traffic charges.
>
> --
>   ++ytti
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/