Our industry has a habit of grabbing ahold of new technologies, slapping them onto marketing visions, and then questioning why things haven’t arrived two years later. These expectations are unrealistic, especially when it comes to building the telco cloud with network functions virtualization (NFV) technology, which is a large architectural leap that will take many years.
If you put NFV into the context of major technology movements, it’s at least a decade-long process. The move from minicomputers to PCs and client/server networking took at least a decade, peaking in the late 1990s. Then there was the Internet, and a bit later, cloud. The shift of enterprise applications from the local networks to centralized data centers started about 15 years ago and is still going on.
Amazon and the Cloud
As an example let’s look at the profile of Amazon’s Amazon Web Services (AWS) business. The revenues are not publicly disclosed, is, but several Wall St. analysts have estimated it to be in the $5-$10 billion range. This business was launched in 2006 — a decade ago. So it took about 10 years to build the infrastructure that could support a multi-billion dollar business.
To take another example, Microsoft launched Azure in 2008, about a year after AWS. It’s been almost a decade, and Microsoft has only recently been acknowledged as starting to compete with Amazon as one of the top three cloud leaders. Google‘s early cloud was built a few years after it launched, when its search volume drove it to scale rapidly. Google went public in 2000, so that’s been over a decade in development, as well.
If you look at the cloud market in general, it’s just come to dominate the technology landscape in the last few years after a decade of development. According to the Northbridge Future of the Cloud Survey, use of public cloud has expanded 43% since 2011 while private cloud has declined by 48%. The survey concludes that public cloud is now in an “acceleration phase” — but this is really 15 years after it first started as a concept.
Telco Cloud Takes Time
Telcos are not known for doing things quickly – they are large organizations with bureaucratic overhead. How would we expect them to build the telco cloud any faster than Amazon? NFV and SDN represent generational shifts in technology architecture. At best, this is a 10-year proposition. The first NFV servers are only just now being installed and tested, after years of initial proofs-of-concept (POCs) and testing. If we argued that NFV really started a couple years ago (which is aggressive, since the technology is only now coming to market), it’s not unreasonable to expect NFV to start hitting full stride around 2019, three years from now.
The North Bridge survey, which is one of the most extensive surveys of cloud businesses I have seen, breaks out other cloud business models that could be compared with NFV. For example, it estimates the infrastructure as a service (IaaS) market is a $25 billion industry with 19 percent compound annual growth rate (CAGR). Platform as a service (PaaS) is $2.3 billion with 38 percent CAGR.
NFV has elements of both IaaS and PaaS, addressing specific communications applications. Many telcos are going to build their own NFV platforms, but it also makes sense that third-parties will build NFV services and clouds that they wholesale to other service providers. If we expect them to be multibillion-dollar markets (which I do), it’s not unreasonable to see them taking the trajectory of IaaS and PaaS markets – that is, taking a decade or so to develop into multibillion-dollar markets.
Complexities to Conquer
The North Bridge survey also points to the top inhibitors of cloud in 2015: security, regulatory, privacy, lock-in, and complexity. It says that all of these are virtually the same as in the 2011 survey, when complexity was added to the mix. As I mentioned, the telco cloud has these same challenges, only in greater intensity, because of the highly regulated and demanding aspects of the public communications market.
So, the next time you are at an industry panel, and those annoying industry analysts and other such antagonists start bemoaning the slow pace of NFV – you can tell them that SDxCentral pointed out that major technology shifts in history have each taken at least a decade.