How an old-school telco gear maker got the cloud religion. Can it convert the carriers?

Telcos feel like they are between a rock and hard place. When you consider the transition to all IP networks, the margin pressures associated with meeting the insatiable demand for mobile data and the threat that over-the-top services represent to their businesses; it’s clear that they are doing more than just trying to change the jet engine mid-flight, they are trying to replace the engine while others are looting the plane for parts. Meanwhile the skies are getting more crowded with more flyers demanding more routes.

Telcos must invest in their infrastructure, even as demand for their services rises. Yet they cannot ask revenue to continue rising at the pace of consumption, and in some cases, such as text messaging and voice calls their revenue is falling. So far their response has been to decry bandwidth hogs, implement new pricing plans that try to hold the line on the dollars coming in even if users choose to use over-the top-alternatives.

airplane_thumbBut some are realizing that that’s not enough. They are investing in technologies such as OpenFlow, or at least software defined networking, as they try to get a handle on their costs. And they are demanding their suppliers provide them with specialized software running on commodity hardware, as opposed to the pricey, proprietary boxes of previous generations’ of technology.

Metaswitch’s big switch for telco gear.

For example, Metaswitch, a three-decade-old company based in San Francisco has created Project Clearwater, a software-based IMS core for telephone networks. An IMS (it stands for IP Multimedia Subsystem for those who care about these things) system is the glue that connects the old analog wireline systems to the newer digital systems. The thought behind IMS was that mobile operators would use them as a bridge into the IP world, but in reality they proved complex and expensive and telcos put off making those investments.

As Metaswitch looked at the market two years ago it saw an opportunity. The company, which provides other hardware to wireless carriers, saw the world was changing. So CTO Martin Taylor said the company decided to build an IMS core that ran on commodity hardware. And if that wasn’t revolutionary enough (remember, we’re talking about telcos here) on May 8 Metaswitch will open source the software.


Taylor points out that telcos used to have the largest scale systems, but that is changing. The globe-spanning networks delivering five nines that once inspired such awe, are now common as Google, Microsoft and others build out their own globe-spanning infrastructures. And telco’s know that to keep up they must adopt the same tricks the web scale companies have, like open source software and commodity hardware. Thus Metaswitch will open source its Clearwater software, and follow a Red Hat model of supporting the software and releasing regular updates. Taylor has the right idea, but telcos need to go even further.

But the real solution isn’t open-source software

As forward-thinking as Metaswitch is with its open source business model and trying to deliver a software-based IMS core built for commodity hardware, its customers are making a mistake if they rely on Metaswitch to hold their hand. As the telco network looks more like cloud and webscale infrastructure — in that telco networks they are taking on more load without adding costs — telcos need to think like real cloud vendors and webscale companies, not like enterprise IT customers.

Telcos are providing essential infrastructure in their mobile networks. Many of them also provide cloud computing services. In yesteryear it was enough to just provide the pipes, but if you’re going to provide compute and networking infrastructure today you need to adjust to the new reality for infrastructure providers.

The new infrastructure reality

223102_commodity_tradingAnd that reality is you need to own your systems. Infrastructure is going to be a commodity, even in mobile access (look at Free Mobile’s plans in France or even Republic Wireless here in the U.S. if you want to see the future). And people are going to want more and more of it, so the build out had better be cheap. So if telcos really want to be cloud providers, and the really want to compete in an all IP world, they need to stop demanding hand-holding from their vendors, hire smart people to own their infrastructure development, and get off their butts and start innovating.

For example, Amazon(s amzn) doesn’t hire a company to provide help on its operating systems or databases. When it chooses an open source technology it also chooses and hires smart people to make sure that technology is up and running and maintained. Google(s goog), Facebook(s fb), Netflix(s nflx), they all operate the same way in the core areas of their business. Because when you cut out the middle man you cut costs. When you have smart people on staff, you can keep innovating at your pace and in the direction you want to go.

So if you’re going to be an infrastructure provider, that mindset and skill set is par for the course. And telcos do not seem to get this.

They can’t say they want to be like Amazon and play in that world if they want their vendors to do the work. They’ve got to find a way to embrace not just the technologies but the economic realities of competing in the commodity and cutthroat business that is the cloud and IP networks. Otherwise they will begin a long decline.