|By Esmeralda Swartz||
|June 29, 2014 09:00 AM EDT||
In my last blog, I looked at the evolution of technology as we move to virtualization. What will this new development mean for business models, industry developments and industry players?
Virtualization means that everything will run on commodity hardware - the same as the rest of the commercial world. No proprietary equipment, just generic devices and servers running software. Who will build the data centers? Probably the same set of companies that build data centers for everyone else. That is, not necessarily network equipment companies or phone companies.
Who will write the software? Presumably the network equipment companies, at first. But the cost of entry into a software-driven industry is much lower than the cost of entry into a hardware-dominated industry. Mass-producing software is far less expensive than mass-producing hardware, so new players will rapidly emerge and some of them will prosper. Just as likely, some of the existing players will not. An industry with thousands of large, medium and small companies has different dynamics than an industry dominated by a few giants.
Will communications service providers (CSPs) get into the software business? They understand how networks work, and the move to virtualization provides a new playing field in which CSPs stand as good a chance as any of creating new, reliable and secure software. Why shouldn't the CSPs work together in a sort of almost-open-source movement, share their products and services and reap the financial benefits?
And then who will design the networks and do the network capacity planning? Ultimately, the network will do most of that itself. Just as it is possible to write code that writes more code according to rules, it is possible to write code within the network that not only manages the traffic, but identifies when and where new capacity is needed, and installs itself in the newly activated devices. Where do those devices and servers come from? Anywhere; they are generic, standardized and commoditized.
Who will create network capacity by actually placing the building blocks in position? Some humans may do this, or maybe robots, but they don't need to work for the phone company. They could work for a sub-contractor or one of the companies that today we think of as equipment vendors. Or perhaps that is a whole new industry, in the same way house builders don't have to make bricks, or be architects, real estate agents or landlords. They just build and sell.
So who will actually own the network assets? It could be anyone, even a company that has no actual end users as customers. Virtualization, and all that goes with it, will make it possible to own network assets, and to offer those assets as capacity for others to rent. That capacity could then be used by whatever CSP needs it most at the time, and is willing to pay for it.
CSPs will both own and rent capacity, or rather their automated agents will do so on their behalf, guided by rules that will be developed initially by human coders, but which will evolve and be optimized by more software. And the fact that all network capability can be addressed by anyone with access means that we could end up with a rich network of niche relationships, with services appearing and disappearing like twinkling stars.
Users of the network today are somewhat imprisoned: they have limited choices of service providers and services. It may be in the interests of CSPs to keep it that way for a while, but ultimately the capabilities of the technology will create a completely new type of customer. The Internet has shown us the model: people can go almost anywhere they want to. But let's not confuse "The Internet" with "The Network." The global telecommunications network is bigger than the Internet, and actually contains it. It is this larger, all-containing global network that is being virtualized by network functions virtualization (NFV) and software-defined networking (SDN). Soon the flexibility and power created by this move to virtualization will spawn all sorts of new services driven by whoever wants them and is willing to pay their share for capacity. The Internet as we know it will be just one particular mode of networking. NFV and SDN can, and probably will, lead to a multiplicity of Internet-like services, both competing and collaborating, with different purposes and philosophies. Some of those exist just now, in the form of commercial, governmental and academic intranets, but we should expect more when anyone can address the network, define the necessary capabilities and use them as needed. These individuals can also sell them to others with any type of charging arrangement: charging for time or for capacity used, by content consumed, or subscription by the hour, the year or the second. And if you need a content delivery network, you can rent space on those built by others, independent of other business relationships you might have. Or you can build one yourself.
So monthly bills? Well, remember the hansom cabriolet? It's cute, but not what we need today.
What other business processes may soon become outdated? Let us know on Twitter.