Nobody doubts that the telecom ecosystem is in need of more innovation. That’s why it’s a good sign that there’s a group of people working to promote just that. They’ve published a paper (available HERE) and they’re inviting comments, so I’m offering mine in this blog. Obviously, I invite the authors and anyone else to comment on my views on LinkedIn, where a link to this will be posted.
The paper is eerily similar to the “Call for Action” paper that, in 2012, kicked off NFV. That the initiative seems to be promoted by Telecom TV also echoes the fact that Light Reading was a big promoter of NFV, including events and even a foundation that planned to do integration/onboarding. These points don’t invalidate what the paper contains, but they do justify a close look.
Let’s then take one. A key question at the end of the paper is “What did we get wrong?”, and so I’ll offer my view of that point first. The biggest problem with the paper is that it’s a suggestion of process reform. It doesn’t propose a specific technology, a specific approach, but instead talks about why innovation is important and how it’s been stifled. With respect, many know the answers to that already, and yet we are where we are.
The paper makes two specific process suggestions; R&D partnerships with vendors and improvements to the standards process. I heartily agree with the latter point; how many times have I said in my blog that telco standards were fatally flawed, mired in formalism, prone to excessive delays, and lacking in participants with the specific skills needed to “softwarize” the telco space.
NFV was softwarization, but it never really developed a carrier-cloud-centric vision. Instead, it focused on universal CPE, and as 5G virtual features became credible, operators started looking at outsourcing that mission to public cloud providers. No smaller, innovative, players there, and in fact really small and innovative players would have a major problem even participating in carrier standards initiatives. I’ve been involved in several, but only when I’d sold some intellectual property to fund my own activity. None of them made me any money, and I suspect that most small and innovative companies would be unable to fund their participation.
I could support the notion of improved R&D partnerships, if we knew what they looked like and what they sought to create. What would the framework for those partnerships be, though? Standards clearly don’t work, so open-source? ONAP was an operator open-source project, and it went wrong for nearly the same technical reasons NFV did. There’s not enough detail to assess whether there’s even a specific goal in mind for these partnerships.
Except for a single statement that I have to add to the “Where did we go wrong” point, with a big exclamation mark. “Softwarization of the infrastructure (i.e. NFV, SDN, cloud) has, in theory, created opportunities for smaller, more innovative players, to participate in the telecommunications supplier ecosystem, but there remain significant barriers to their participation, including the systemic ones identified above.” It’s clear from this statement, and other statements in the paper, that the goal here is to improve the operators’ ability to profit from their current network role. Innovation in connection services, in short, is the objective.
I’ve said this before, but let me say it here with emphasis. There is no way that any enhancements to connectivity services, created through optimization of data network infrastructure, can help operators enhance their profitability in the long term. Bits are the ultimate commodity, and as long as that’s what you sell, you’re fighting a losing battle against becoming dirt. Value is created by what the network supports in the way of experiences, not by how that stuff is pushed around. Yes, it’s necessary that there be bit-pushers, but the role will never be any more valuable than plumbing. The best you can hope for is to remain invisible.
As long as telcos continue to hunker down on the old familiar connection services and try to somehow make them highly differentiable, highly profitable, there’s nothing innovation can do except reduce cost. How much cost reduction can be wrung out? Cost management, folks, vanishes to a point. There is no near-term limit to the potential revenue growth associated with new and valuable experiences. The moral to that is clear, at least to me. Get into the experience business.
Smaller, more innovative, players have tried to break into telecom for at least two decades, with data-network equipment that would threaten the incumbents. It didn’t work; the giants of 20 years ago are the giants of today, with allowances made for M&A. The places where smaller players have actually butted heads with incumbents and been successful have been above basic bit-pushing. Trying to reignite the network equipment innovation war that small companies already lost, by moving the battlefield to software, just invites another generation of startup failures. Anyway, what VC would fund it when they can fund a social network or e-commerce company?
Innovation, meaning innovators, go where the rewards are highest. Telco-centric initiatives are not that space, as VC interest demonstrates. Will the operators fund initiatives to improve their infrastructure? I don’t think so; certainly none have indicated to me that they’re interested, and without some financial incentives for those innovators, innovation will continue to be in short supply.
If operators are serious about bringing more innovation into their operations, they need to start by setting innovative service goals. Different ways to push bits are really not all that different from each other. Different ways of making money off bits is a totally different thing. The whole of the Internet and the OTT industries were founded on that, so it’s perhaps the most proven approach to enhancing revenue we could find out there.
There is value to things like NFV hosting and 5G Core virtualization, as drivers to an expanded carrier cloud infrastructure that then becomes a resource for the other future higher-layer services an operator might want to provide. However, we don’t need innovation initiatives to do this sort of thing. We have cloud technology that’s evolving at a dazzling pace and addressing all the real issues. I’ve blogged in the past about how cloud-native control-plane behavior, combined with white-box switching, could address connection-network requirements fully. We don’t need to be able to spin up routers anywhere, we just need to spin them up at the termination of trunks. But even white boxes can be part of a resource pool, supported by current cloud technologies like Kubernetes.
We don’t need NFV, nor do we need an NFV successor. We can now turn to the cloud community for the solution to all our problems, even problems that frankly the operators don’t even know they have yet. Where innovation is needed is in the area of service planning. Bits are ones and zeros, so differentiating yourself with them is a wasted exercise. It’s what they can carry that matters, and so innovative operators need to look there, and leave cloud infrastructure to people who are already experts.