The recent latency issues are not on MTN's side, they were actually due to us (afrihost).
We've allowed all users (capped and uncapped) to run pretty much unshaped since change over so that we could ensure that even in a worst case scenario we a) had enough IPC capacity (which is being upgraded anyway) and b) that the network would continue to perform under stress. While the network help up it's end, without proper shaping policies in place, we had the problems of latency and ping causing some users issues (while most were getting full performance). We should have implemented shaping and QoS yesterday, but for some reason this did not take effect in time ( I am still waiting for answers on this) and I have been assured by Gian and the other Directors that this will be in place by end of business today.
Usage spikes just after 6pm, which is when things have been going pear shaped, and it's mainly because uncapped accounts were basically performing at the level of capped accounts - no shaping, no throttle, no prioritisation. We're tweaking the uncapped accounts now to get the best overall performance, while recognising the priority of capped users who pay per GB. We're trying to give everyone the best overall experience when usage is high, and also a premium performance for uncapped users when we can afford to give them more capacity/resources.
Our network is 17 days old (in terms of being live). It takes some time to get things fine tuned for best performance. If we look back just a week, there's no doubt that performance is getting better and better for more clients as we progress. It's a journey, and we fully plan to reward clients who stick with us with better pricing and products in the future.
