Re: From a technical point of view, a single network makes sense
Hi Dave. Interesting post and a couple of observations:
1) On the issue of spectrum - The maximum channel bandwidth in LTE is 20MHz. Whether you stack 5 x 20 Mhz channels adjacent, each still loses about 10% due to guarding, regardless of whether they belong to one operator. There is no way LTE allows for a contiguous 100MHz block. Only in 5G can contiguous blocks of this size be realized. The spectral efficiency between 5, 10, 15 and 20MHz bands differ very little. Only in 3Mhz and lower do the overheads really become a factor.
2) Massive MIMO is a very new, mostly 5G tehcnology. If you look at the technology distribution in African networks you find that more than 70% of subscribers still use basic feature phones. The ones that use LTE smartphones are usually entry level, Chinese whitelabelled models. The percentage of devices that support the very latest in radio and antenna technology wil most proabably not even number 1 % - it is simply a question of affordability. This means that building networks using the very latest wizardry to mitigate a spectrum shoprtage will do very little to relieve congestion. The only other option is to densify the network with more sites which eventually gets transfered to the subscriber.
I sincerely doubt that vendors can tell you that Massive MIMO is rolled out in the thousands. Even 8x8 MIMO used by Huawei in offerings such as WTTx has only been around for about 18 months. And Massive MIMO defintely does not increase throughput 3 x 10 fold. That is once again vendors slideware :-)
Many South African towers are shared towers meaning they have 2 or even 3 operators on them. It is not simply a question of adding more antenna elements. Space is a huge factor on the towers. Already most oeprators have multiband atennas spanning 900/1800/2100. Also MIMO only really works in urban environments where one can benefit from multipath. The real problem faced by SA operators is that no sub 1GHz spectrum is available for LTE. This means it is very costly to deliver in-building coverage and simply impossible to roll out LTE to rural markets using refarmed 1800MHz.
A profitable mobile network sweats its assets over multiple revenue streams such as voice, data, Mobile Money, Digital Entertainement etc. There is simply no way that a greenfield network that has no legacy revenue will be able to fund the rollout of a new nationwide network which will eventually pay for itself whilst at the same time delivering data at a wholesale cost which is lower than than of the current operators. If you need convincing look to Rwanda where the wholesale cost of data on the monopoly LTE network is 30% more expensive than the other operators' RETAIL 3G data prices. After 2 years they have managed to connect less than 10,000 subcribers. There is simply no way operators can compete purely on a service level in the wireless space. Your pricing as well as customer satisfaction is absolutely related to the technical strategy with whcih you run your network.