Will software eat MWC?

It became a running joke, a partially-funny metric. How long could you have a conversation at MWC without mentioning the $19 Billion Facebook acquisition of WhatsApp? If you made it through an entire hour meeting without the subject coming up, you were probably talking about semiconductors, or optical fibers. In almost any other conversation, that $19 billion was almost guaranteed to come up.

You can argue that Facebook overpaid, but it is hard to argue with the simple fact that WhatsApp, with its 30 engineers, built a system capable of handling a volume of messages equivalent to the world’s entire SMS text messaging traffic. There is a great post on High Scalability looking at WhatsApp’s network architecture. Put simply, they had a very experienced team using some very inexpensive software to design massive scalability. (On a side note, it sounds like WhatsApp hosted its own server, one more start-up not built on AWS.)

For companies that build “carrier grade” SMS-handling network appliances, you have to be very concerned. And there are a lot of these companies at MWC every year.

I did not have enough time to really dive into the details with the big equipment makers this year, but even in the little time I did have, I sensed a healthy level of concern. The official marketing position remains something like “WhatsApp is a special case, and if you need something that is Five 9’s reliable, and capable of meeting regulatory requirements, WhatsApp’s approach would not work.” But peel back the marketing and there is considerably less substance to their defense.

WhatsApp is pretty reliable, and seems to be built using the Erlang programming language (which was developed by an equipment major), and is known for being highly reliable.

Moreover, the price tag for WhatsApp really calls into question that marketing position. And more than that, it calls into question what business the carriers should be in. It has been clear for a number of years that the carriers know their messaging revenue is in permanent decline. To handle that, they began by bundling voice and SMS. Then, with voice minutes also declining, they began to bundle voice and SMS with data. The wireless operators are very good at segmenting users and pricing to each user niche. But what can they bundle next? Why not just admit that everything is now ‘over the top’ and focus on data pricing alone.

It should be clear now where the technology stack is headed, and many carrier ‘services’ are becoming apps, or just features in apps. Dean Bubley, for one, has been talking about this for years. With WhatsApp, you can see the various carrier defenses laid bare. There will always be value to providing ubiquitous access (i.e. cellular base stations and last mile broadband), and someone will need to handle long haul transport. These are very lucrative businesses, but the many other features and services that rest on top of it look severely challenged.

The carriers have their barriers to entry – regulatory, political, zoning, and real estate, but for the equipment vendors who provide those various services, the outlook is much less upbeat. The equipment majors have all started talking a lot more about their software and services platforms in recent years. They have the scale to adapt and acquire entire new capabilities. But there are many point specialists, selling single-purpose hardware appliances or near-custom software solutions for carrier services. Once upon a time, this was a highly profitable business for a whole swathe of the industry. That position now looks very precarious.

Leave a Reply