Recently I revisited a blog I wrote in 2015 (see below), prompted by Appledore’s attendance at SDN and NFV World Congress (see NFV Market at a Crossroad)
Deja Vous all over again
What surprised me in re-reading this was how little has changed in our industry in 3 years. 5G still seems to be primarily about radio technology upgrade. NFV, where it is being deployed, is largely being used to replicate existing physical network infrastructures or creating limited evolutions of connectivity services and being deployed into the existing manually intensive operational environment. Appledore believe that this limited innovation and limited automation is meaning that the benefits of software enabled networks is not being realised. It is creating an environment in which investment in software enabled networks (beyond a network replacement technology) is being questioned. Appledore explores this issue and possible remedies for this in our new market outlook “Cloud Native – a revolution postponed”
https://appledoreresearch.com/report/cloud-native-a-revolution-postponed/
Why NFV should Lose the N
Telecom’s positioning of NFV (Network Function Virtualization) as a distinct concept distracts from the real opportunities in virtualization. It will ultimately limit the applicability of these network functions to simply replicating what existing networks do – and miss the real potential.
NFV is just one more example of the virtualization of functions throughout IT. By treating it as a separate concept we inherently segment what is network function from software functions – thereby removing the easy opportunity for the combinations of these in novel services. It’s time to lose the N!
A driver behind this separation is the belief that telecommunications requirements are in some way distinct and more stringent than others; the five nines mantra! Alan Quayle in his post from 2014 nicely sums up the fallacy of this.
“Of course, the telecoms industry thinks it’s special with respect to its requirements. It’s not. Telecom is not the most reliable, it doesn’t deal with the largest data set, or provide the most secure environment. It doesn’t require the lowest latency, etc. Financial services, military, government, healthcare all have more stringent requirements.”
We need to recognise that existing network architectures are simply IT applications when they become virtualized. NFV projects, such as the Clearwater IMS open source project, are interesting but inherently limited to replicating an existing physical network architecture. We have to question whether simply virtualizing an unsuccessful telco application (network architecture) changes the willingness of others to use it. It does not change the lack of take up of this technology from both telcos and those outside telcos.
Increasingly we need to think about how network functions can be unbundled in a way that can be used by any IT developer to create innovative services. Many adjacent industries require the capabilities of the telecom network but do not necessarily want the whole solution and business model of today’s telco. They want to be able to use telco capability and build it into their own business models. A change of approach could see others valuing the telco, rather than doing their damnedest to avoid it for anything beyond base connectivity.
5G offers a tantalizing fork in the road. Looked at one way, it’s just another radio access technology, with all of the attendant processes issues that implies. But looked at from the perspective of an IT function approach, there’s a completely new outcome, enabling 5G to become a valued business capability relevant to multiple industries.