Commoditisation: Much needed in the service enabling layer
At the recent WDM and Next Generation Optical Networking event in Monaco, it seemed that all of the optical elements that were discussed are effectively perceived as merely network commodities. They all do more or less the same things, just faster, bigger or better. Certainly the operator speakers seemed to be in agreement that the cost of these must be driven ever lower, which is certainly characteristic of a commodity driven market. More on this later.
At RCL, we normally work with vendors that operate at the application layer, and here the question of value is important. Typically, in any tender for an application with a service provider, a significant proportion of the cost is consumed by the underlying technology, principally the application server. Now if every deal includes an application server, one is forced to conclude that the cost of every application deployment ends up being much higher than it perhaps should be. The question is, does this inhibit the market?
In the shiny new NGN world, the application server is supposed to be an all-purpose piece of kits that sits above the network and presents suitable interfaces to applications; in the IMS world, connecting to the ISC interface. In other words, while it adds value, it is supposed to be the key to unlocking a stream of easy-to-deploy applications which actually generate cash for the operator. Just read the 3GPP specifications to see how that was the vision. That is, it sounds very much like it should be more of a commodity item, just as, arguably so should the media server (MRF) in the same architecture.
But this hasn't happened: could it be that the fact that the cost of application servers remains so high that has held back deployment of innovative applications? There are lots of innovative applications around - our customers Gintel, NetDev and Almira all have interesting offers and there are many more besides - but the overall cost of new application delivery is inflated by the need to purchase application servers from a limited number of vendors. Isn't it time their potential to host multiple applications was really leveraged so new applications can be added easily to an existing infrastructure, just as was originally intended? After all, one of the promises of IMS was that the incremental cost for deploying additional applications would be lowered as they would all leverage a common underlying architecture, including the application servers.
Of course, we recognise the particular value that application servers offer, but is that value really equivalent to the applications which actually generate revenue? Driving down the cost of the application server and turning it into more of a commodity would certainly help service providers deploy more applications and, in turn, help nurture an expanding ecosystem of application developers whose market entry costs are significantly lowered. Who knows, maybe even more application servers would be sold as a result?
If we want innovation in the telecoms industry, wouldn't it be useful to turn to a willing crowd of knowledgeable application developers, familiar with the challenges faced by operators, rather than trusting in an unproven and extremely risky application store model?