I am constantly thinking about just how network functions are going to become virtualized effectively enough to impact carrier transport networks, where you see the need for very large bandwidth circuits. The current state of virtualized network functions (VNF) is such that you wouldn’t even consider going down this path due to the limitations that exist today. The coming containerization of VNFs is not going to fix this issue either. What might fix it? Well, believe it or not, quantum computing.
Before we decide I have gone off the deep end, let’s do a few things like define just what that means. I was reading an article over the weekend that was talking about just how close quantum computing is to becoming a reality in more than just a lab setting, and it struck me how the benefits expected could directly apply to the future we are all working toward right now. The best explanation for quantum computing is best left to an expert, so I’ll leave that to the following video:
There is a concept of “quantum supremacy” that basically means the machine being tested against the standard (e.g. a computer/server) is indeed a quantum computer. Two major organizations working on this, Rigetti Computing and Google, believe they will reach quantum supremacy within the next 12-18 months, so this is not science fiction. This is real. To learn more about this in a wider context read the article that I obtained the video above from.
In short, the video explanation, for those of you that chose not to watch, boils down to this. The actual transactions that take place in the computer will not speed up in any significant way. There are limitations of the hardware involved that limit this. Those transactions will be just as slow/fast as they are today. However, the amount of information that can be packed into each transaction will increase exponentially. New operations are being defined and implemented to leverage this ability to pack massive amounts of information into each transaction. As a result the number of transactions required to achieve the same results are expected to radically decrease. Radically decreased numbers of transactions, even if each one is as slow as today, equals radically faster processing of information overall. This has direct implications on NFV in the areas of VNFs, in the area of SDN controllers that will need to dynamically adjust networks based on certain conditions, and even directly within OSS when it comes to things like “big data” for predictive analytics and automated healing of networks.
Bottom line? Current computing technology, and the rate of improvement, is not quite enough for us to reach parity between physical and virtual devices…today. As the architectural design of those VNFs improves, current technology improves, and new computational technologies, such as quantum computing, are introduced we will see these more advanced forms of virtualized networks become reality.
I think we will see things like virtual collocation spaces, where a carrier will log into a portal when they have a need to colocate equipment in another carrier’s location. All carriers will have NFV infrastructure in place in each of these locations available for that portal to provision an NFV matching the specs required, including carrier class bandwidth requirements for transport. That is the future we are headed towards, and it is technology like quantum computing that will allow us to get there. To achieve this is going to require a robust, and dynamic, management ecosystem…but we talk about that all the time…