[Afpif] Transit vs. peering Focust for Africa by 2021

[Mark Tinka]

On 6/Aug/15 18:04, Meoli Kashorda wrote:
> Dear all,
>
> 80% local traffic is very ambitious if that means locally generated
> content. But if it means hosted in / served from Africa, then we could
> achieve that in the next 2 year

Agree.

> - I am not surprised by the 60% figure. It is possible we are already
> at 50% in Kenya because of Google and Akamai caches. I would like to
> see the per country figures. Unfortunately, Most of this traffic is
> not visible at the IXPs because Google caches are in different
> networks and private peering.

Unlikely.

I'd say the edge caches you mention account for single-digit proportions
of overall traffic within all African countries. Yes, the traffic could
be significant in the largest of networks, but it is still a fraction of
overall demand.

There is still a lot of traffic that is not physically available in
Africa. And that is the over 90% we all pay dearly for to go fetch all
the way from Europe. The big names currently in Africa are by no means
the lot; they are just the bold ones.

>
> However, It is also important to unbundle local Internet traffic -
> most of the traffic could be video and YouTube traffic but the
> relatively low traffic generated by e-commerce, egovernment, email or
> E-learning applications, and local websites might be what really
> drives the Internet economy in our countries.

I'm not too concerned about what type of traffic will turn the tide, as
long as the source and destination of that traffic remains in Africa.

Mark.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.afpif.org/pipermail/afpif/attachments/20150806/8531b6c7/attachment-0001.html>