It’s All Too Good To Keep Talking About The Capacity Problem
October 11, 2014
The capacity problem is at the heart of everything said about the wireless industry. Everybody loves to talk about this problem. To start, it is easy to give examples of exploding data consumption forecasts or quote numbers on mobile applications such as Facebook, Twitter , Instagram and many others. I suspect that the capacity problem makes for a convenient argument for the different players in the mobile value chain to get what they want: it is a nice problem for everyone to have. For service provider side, it is the key to more frequency spectrum which further enhances increases their position in the market and consequently their value. For solution vendors it gives them the opportunity to raise money, fund development projects and present forecasts for high revenues. For regulators it gives them the opportunity to sell spectrum and raise money. So everyone can benefit from the capacity problem, or so it seems.
The one category that would have much to lose from the capacity problem is investors. Billions of dollars were lost in WiMAX because 3G networks were not loaded and required upgrade (there are many other reasons for WiMAX’s failure, of course). Similarly, hundreds of millions were lost in the femto cell market (femto cells have been deployed for coverage but not capacity reasons – volumes are predicated on capacity use case that never materialized). Today, we see the same thing repeating in other segments such as small cells where every year we hear forecasts for deployments ‘next year.’ The fact is, wireless networks are not fully loaded as is telling by the number of LTE subscribers and new spectrum allocations where new networks are in process of getting deployed (LTE provides higher capacity and spectral efficiency than 3G, so it makes sense to load LTE and migrate as much 3G subscribers to LTE before looking at increasing the capacity of networks in a way that requires heavy capital and operational expenditures). But aside from the loading of wireless networks, I think there’s a fundamental change in how capacity is offered and consumed that makes the value proposition for many ‘capacity solutions’ flawed – at least for the time being.
When talking about capacity, many still think along the lines of busy hour traffic on circuit-switched networks which predicates many ‘running out of capacity’ forecasts. This approach of thinking about capacity is logical and straight forward. It is how engineers plan network capacity even today when networks have changed from telephone networks for voice service to data networks for Internet services. However, data services present a whole new set of tools to operators to control traffic on the network. Consider for example the tests done by Ericsson and SoftBank in implementing mobility-based policy management by directly connecting the SGSN-MME to a service-aware policy controller which allows dynamic allocation of resources on per-user basis. This technique can be carried further to be able to control resources on an application basis within different context such as time of day and location. So in short, there are real alternatives to traffic management that are yet to be fully commercialized that will further extend the capabilities of wireless networks. The availability of these options will only make planning for capacity more complex as predictability of capacity gets fuzzier, which in turn will obscure to the wireless ecosystem a key parameter that operators guard closely. In other words, even as new and innovative capacity control mechanisms will be implemented there will be ever more talk of running out of capacity! It only bodes well for all parties.