It’s apparently time to say goodbye to Metaswitch, one of the most potentially transformational telecom vendors/startups of our time. The company was a provider of open-source telco software designed to implement the features of the IP Multimedia Subsystem (IMS), and a key part of the multi-vendor alliance that I assembled to provide a cloud-centric implementation of the then-emerging concept of Network Function Virtualization (NFV). Light Reading (who offered this view on Microsoft’s decision to drop the telco tools it got by acquiring Metaswitch, named Martin Taylor, a Metaswitch executive, as one of the most influential players in NFV. What happened here?
The two most obvious things that happened was the acquisition of Metaswitch by Microsoft, and the fact that the open-software telco infrastructure revolution it represented could be inspired by a startup but completed only by one of the telecom sector’s incumbent vendors. I think that these two obvious points are related.
The most significant factor in the evolution of telecommunications is the “IP dialtone” phenomena. Today, every meaningful new data service that has ever developed is created “over the top” of the Internet, and it seems highly unlikely that will ever change. It’s also true that this dialtone phenomena has shifted the focus of pre-Internet (or at least pre-Internet-dominance) service implementation toward that same IP dialtone overlay model.
This combination truth also leads to the next issue, which is the displacement of traditional services by OTT and mobile apps. Telephony has inter-calling and cross-messaging standards that are critical, and whose implementation required some sort of standards set like mobile’s IMS. The broad group of OTT apps we tend to call “social media” has rendered a lot of this obsolete. Instead we have opt-in social communities that have their own common client apps, and interoperation isn’t a factor. Traditional calling is for many more an intrusion than something to be cultivated, and while texting does require some interoperability among services, it is falling out of favor as a regular means of interaction within a social group. In all, OTT applications, even simple email, are displacing more and more “telco-type” services, which limit the value of something like IMS. Revolution has overcome evolution, in terms of telecom infrastructure.
I also think there’s a healthy dose of reliance on hype here. 5G generated a lot of media attention, but nearly everything said about it turned out to be exaggeration at the least, and often simply wrong. Did Microsoft think they could use open-software technology and cloud computing to displace major incumbent vendors on big deals, or were they satisfied with lower-tier telcos?
Maybe they believed in NFV, though Microsoft wasn’t a mover in that group, because Metaswitch surely was. The goals of NFV, which were involved use of virtual functions to displace appliance dependence, were surely compatible with Microsoft’s vision, and it should have been clear to everyone that cloud technology should have been the explicit foundation of NFV. But NFV never really addressed the foundational points in function hosting, and so it was not likely to drive any adoption of the model Microsoft might have hoped for.
If you look back to packet switching (X.25/X.75), the “Integrated Services Digital Network” (ISDN) and Asynchronous Transfer Mode (ATM) you can see the same sort of thing at work. In all these cases, the telecom space was responding to an emerging market opportunity, in an evolutionary way. In all cases, they were unsuccessful, in the main because they “shot behind the duck”, by being too late overall or failing to anticipate the real basis of the opportunity. The market will always seek the largest profit and ROI possible, as fast as it can be delivered. One OTT visionary told me over 20 years ago that “first-mover advantage is the only advantage that counts.” Telcos rely on standards, and standards never deliver any of that. Where, in any of the standards I’ve cited or in any of the mobile standards, do we see a real analysis of the business cases that will generate the benefits, the profits?
The notion of IP dialtone disconnected telecom services from retail services, forever. The profit-driven innovation of the OTT space has run with the notion that everything in the future should be over IP, and that IP is all that telecom should provide. This is not a perfect solution because the division between providing the dialtone and the service has divided us into a profitable space and one that’s increasingly not profitable, but is absolutely essential. How that’s resolved is yet to be known, but it must be.
The fault here, though, lies with the telcos. Orderly evolution is the goal of telco standards processes, which is understandable at first but unreasonable if you look deeper. The cost of change, even revolutionary change, is only one ingredient in ROI, the “I” piece. If you have enough “R”, then revolution is actually a good idea. The fact is that the goal of telecom infrastructure has to be revolutionary, while at the same time not forcing unnecessary or unjustified displacement costs. IMS should have been what the name suggested, a way of making multimedia and mobile infrastructure converge, not a way to accommodate the thing that should have been recognized as the primary traffic driver—content—with minimal impact on mobile. Mobile needed to have a minimal impact on content.
Telcos I talk with blame regulators for this. They point out that telecommunications as we have it today evolved from either a public-utility model or from an entity that was actually part of the government (Postal, Telephone, and Telegraph or PTT). It’s still held in by antiquated regulations on one hand and bound on the other by politically volatile net neutrality policies.
I think regulation of telecom has been ineffective, in some cases for political reasons but in all cases because regulators lack the technical background they need, and the economic advice that has to underpin any successful regulation of a major market. But in the end it’s the telcos themselves that have to set the agenda for their business, to educate and influence policy-making. OTTs have done that, and telcos have been largely unsuccessful in setting their own agenda.
Metaswitch was the right technology for its time, which was in the 2010-2014 period when it could have been possible to do NFV and telco evolution in general in an appropriate way. It was simply overtaken by events, or maybe the lack of events.