Thomas Howe

Last month, an article I wrote for IPConvergence.tv birthed quite a spirited conversation on the blog, and I think there’s another way of making the same argument that I wanted to put forward.

The main point of the article was that IMS was a heavy weight architecture, aimed at large carriers interested in large scale deployments with high levels of control. In my experience, innovation comes from light weight architectures, aimed at smaller service providers interested in niche deployments with high levels of customization for those downstream from them. Essentially, if the main point driving IMS was in support of innovative services, I believe it that is has failed in achieving it, as it makes all the wrong design decisions.

Here’s another way of making the argument. In everyday life, there’s a correlation between feature set and mass adoption, and that is the simpler the feature set, the larger the audience it can serve. Pure commodities like corn, water and sugar are universally purchased and consumed, and are in fact prized for their homogeneity. As a farmer, you don’t want your corn to smell or taste any different than somebody else’s. Economies of scale work well for these products, and in the book The Omnivore’s Dilemma, you learn that 25% of all products in the grocery store are derived from corn. As you go up the feature complexity line, and let’s keep it with food, you see that bread is again fairly generic, but now there’s a number of different sorts of bread to satisfy customer needs. Sandwich bread is different than dinner rolls, croissants are quite different than blueberry muffins. Not only do the features become more important and numerous, but the apparent audience is less for each one. Finally, any bake shop or patisserie shows you the wonderful complexity of presentation, color, taste, smell, purpose and audience. The very same raw ingredients are used to make clothes, bags, perfumes, additives, food coloring, artificial flavors… the list goes on forever. Here we see niches in full bloom, where the end results of complex features are not only (much) higher margins, but smaller audiences and more emphasis placed on customization and hand crafting.

Does IMS really support the carriers in providing this raw material to be used by others in innovative ways? It could, and I hope it will be. Carriers like BT and Vodafone get it. Unfortunately, the remainder of the large carrier world is dominated by those that that believe that it’s the carriers job to provide the ten thousand products, and not the homogenous corn. You can understand why they want this to be true, as the margins on wedding cakes are insane. The basic trouble, though, is simple: nobody would eat wedding cake every day - and that’s not a big enough business for the carriers, even if they could make a decent one. IMS deployed by carriers to support other’s applications makes a lot of sense, but requires the carriers to come to terms with their position in the food chain. Many people disagree with my view, including many whom I respect and whose technical skills I trust and admire. Honestly, though, it’s hard for me to taste it differently.