O3B Details Plan for Satellite-Based Bandwidth For Africa

slash-sa writes "O3B Networks has been quietly preparing itself over the last 12 months for the moment last week when it announced that it was going to be offering cheap, low-latency satellite bandwidth that can cover any part of Africa by 2010. It has put in place early finance with Google, Liberty Global and HSBC. Here are more details from the entrepreneur behind the project, Greg Wyler."

Comments

Impractical.

[Whiney Mac Fanboy]

Should've just dragged fiber.

Do you have any idea how the logistic problems with trying to lay fibre from Satelites down to Africa?

First you've got to fire your rocket carrying the fibre up, get it to loop over the satellite without destroying it, then have the rocket carefully navigate back to your chosen destination.

No - wireless is a better solution with satellites IMHO.

Harder then it seems, trust me

[jiggerdot]

I currently work for a company providing IP communications via satellite (both inclined and geostationary). Most of our customers are in Africa, and include some of the biggest ISPs in the more developed regions. since the bandwidth market there has been exploding in the last several years. So I know what I'm talking about when I say this guy sounds VERY optimistic.

The idea of using low earth orbit satellites is great as the latency on geostationary is indeed horrible. you're looking at a minimum of 500ms just to reach the ISP installation (in the US and Europe, in our case) and the RTT to your destination on top of that.If you run into another satellite link on the way, that's 1000ms minimum. so 123ms sounds terrific. BUT:

1) The guy flippantly says "If they want a gigabit, we'll give them a gigabit". For a gigabit, you'll need to work several transponders, with some insane modulation scheme (highest practical I've seen is 16psk, they'll need something MUCH more dense). The higher they go, the more error prone they get.

2) LEO will require tracking, or very high power. which means either a very powerful HPA (for the small links - the ones without the 3.5 meter dish) or a very fast tracking system for the large links with the dish. And what happens when you have to switch satellites?

3) They're looking to solve the last-mile issue with WiMax. This will interfere with C-band transmissions, so I'm assuming they will go with Ku-Band or higher, which is extremll sensitive to rain fade. Africa has quite a lot of rain. Combine this with point no. 1, and you're in trouble.

4) The article indicates they will give the customer a VAST or transmission station and all is good. It is not. Africa is not a nice place. equipment gets stolen and sabotaged. This is from sad experience. And if you do not have techs on the ground (which are very hard to find, at least competent ones) you're stuck either telling the customer "sucks to be you" or trying to support him through the phone with the replacement of a transmitter, which is a bit like trying to help someone fix an engine by correspondence.

5) The human factor - Without sounding too patronizing, the guys in Africa (even the more professional ones) need a LOT of hand holding. I truly hope they have a big and competent support department and NOC staff at the ready, who can understand garbled English through a bad phone connection, as these guys will want help with everything. From helping to identify which device in the network is causing congestion on the link, to "IP experts" who will be brought in to bring up a BGP session and will not know how to access the router, and will want your help in resetting the password step-by-step. You can, of course, tell them to manage their own networks, but you WILL lose customers. That's a lesson we learned the hard way.

In short, good luck to them, but if they truly think the technical challenges are the only ones, they're in for a very nasty surprise.