[kictanet] IG Discussion 2009, Day 5 of 10- Criticical Internet Resources, IXPs and NOFB

Michuki Mwangi michuki at swiftkenya.com
Fri May 1 12:47:07 EAT 2009


Hi Walu, et al,

The KIXP has had a great impact in Internet scene in Kenya.
There are currently 27 peering members with aggregate traffic during
peak times averaging 45mbps.

Question: is this optimal?

I would say that this is not optimal utilization in view of the overall
total transit (whats not passing at KIXP) capacity in Kenya. Therefore
as you rightfully put it the greater percentage of our traffic is
International

Whats its use then?;

Well for a start, this year i have to make my annual returns online and
its pretty fast from my home connection. The Safaricom IPO had similar
experiences. Most importantly is the ability to create a stable local
internet infrastructure. It means that life can go on, should the
satellites and cables go down (and they will).

Whats missing;

There are some missing items, as mentioned in previous emails which are
dependencies.

1) With increased penetration (no of subscribers) we are bound to see
this reaching 100mbs and higher - hence more affordable access for
non-corporate users and more residential users online.

2) local content hosting - It means we need more collo facilities built.
This will also attract CDNs who are keen to do edge caching and require
reliable hosting services.

It also means that theres a market opportunity for collo hosting
services which is yet to be tapped - see my previous emails.

3) E-government - you will be surprised to note that KRA has more
traffic at the KIXP than some well known service providers. It therefore
means that we need more e-government traffic

4) This leads to the next point of creativity on relevant content.
E-government content is one, we need alot more of Free SMS website (its
a pity sasanet collapsed it had considerable traffic at the KIXP.
E-learning is something that would attract home users to get online
off-peak hours. Basically e-commerce as it were.

On this point, i find it interesting that, we have skipped web content
and moved directly to mobile content. It may require that we make steps
to engineer mobile content solutions to work for the web too.

on NFOB;

The Government approach must be appreciated. I was also glad to hear
that now KPLC has been issued with the license to sell the extra
capacity on its fiber infrastructure. Can we have more of this please.

I cant wait for Kenya pipeline, Kenya Railways and others take the same
approach. The more cable we have the lower the access costs.

Can Kenya power, TKL and city/municipal councils lease out sections of
their poles to fiber/cable infrastructure builders please. If we dont
cease the opportunities to build the last mile in respect to telco's
concerns the we cannot achieve complete access.

If there alot of fiber on the ground - it will not really matter who
runs it; the person who runs it will be the one who can raise the
highest revenue for the investor in a highly competitive market space.
All sorts of models will come up and interesting ones too.


Walu, its not possible for such an infrastructure to replace an IXP.

Even if dark fiber was dirt cheap, its a logistical nightmare and
financially unrealistic  to maintain close to 30 connections to each
provider you want to exchange (peer) traffic with. Buts its financially
viable to have one connection to a single location where you connect to
everyone via a layer 2 switch. In this case, you only pay for the half
of your circuit to the IXP location and manage one link like everyone else.

Regards,

Michuki.




More information about the KICTANet mailing list