AndrewAlston
Group Head of IP Strategy – Liquid Telecommunicati
<tinfoilhat>
MWeb sees they can't pay their bills at the end of the month. They decide to introduce severe shaping and feign instability to reduce overall bandwidth usage, effectively not using more than their purchased capacity. This in turns saves their business while screwing their customers in the process.</tinfoilhat>
*Sigh*
I've tried to explain this to people so many times... redundancy can be built to take away the reliance on third parties... but someone has to pay for it...
The subscribers scream and cry for cheap uncapped broadband, but they don't accept that redundancy and cheap don't come together.
You can have amazingly redundant service, or you can have a service that is affordable with as much redundancy as is possible for the price being paid, you cannot have both.
Let's talk for a second about what the redundancy you are referring to would cost...
If you are talking about a dark fiber route back to Johannesburg to give protection against Neotel failures... you are looking at a fiber build cost of *AT MINIMUM* R60 million, plus another approximately R20 million for the hardware to light fiber over that distance, plus the OPEX cost of running that equipment and that fiber. Someone has to pay for that.
If you are talking about redundant SAT-3 capacity... ever wonder why because SEACOM there was no uncapped? SAIX pricing is completely un-affordable at the speeds required for an uncapped service.
If you want to know why providers can't afford to put in redundancy like this, look to the REAL source of the problem, the absolutely insane prices charged by Telkom for access to the DSL Cloud (so called IPC bandwidth), if that wasn't costing what it did, ISP's could afford to build in a LOT more redundancy. Keep in mind, it costs more to serve a DSL customer over IPC than it does to take traffic from London to Johannesburg over SEACOM. And not a little bit more, like, multiple times the amount...
Just something to think about...