The fusion of the telecoms sector with the wider tech industry is inevitable

We spoke to Dennis Hoffman, SVP and GM, Dell Technologies Telecom Systems Business (pictured), who thinks with the proliferation of Open RAN and edge computing, a fusion of telecoms and the wider tech market ‘has to happen.’

Open RAN is something that’s been talked about for a long time, before even there was anything to point to. You have lots of people in the telco industry quite passionate as to whether they are for or against it, and I wondered how you see the current state of the technology from a position outside of the core telco sector.

We obviously have to have a perspective on the industry given who we are and what we do. As the world’s largest diversified computer company, the notion of running applications on open infrastructure, and increasingly defining that infrastructure with software through things like virtualization, is a long, long standing trend that has gone through the IT data centre and has helped spawned the public cloud.

We see a lot of it leaking, frankly, from the IT data centre of the telco into their network core, the majority of which are already virtualized and open. Open RAN is just a logical extension of saying ‘hey, this opening trend is just going to keep on rippling through the network inevitably.’ For a company like Dell, it creates great opportunity because all of those open systems and open infrastructures, whether they’re the core of the telco and/or the RAN, run out of stuff we make all the time.

So for us, a couple of years ago we decided [the] very, very interesting telecom market was coming our way architecturally and we needed to become a lot more focused on it. So we created a telecoms division within the company to do that, and I was asked to lead that group. And we’ve been at it for roughly two and a half years now. Where Open RAN is happening, generally we’ve been we’ve been selected and we’re involved. Most notably close to home [for the UK] is the Vodafone trials.

And yet, I think the way we would look at it is: RAN, the actual act of being a radio network is a very high performance application. And most high performance applications take a while to work as well on an open disaggregated software defined infrastructure as they do on closed proprietary hardware. And that’s the debate I think you’re seeing with passionate viewpoints on both sides. Those whose business is very much tied to the closed proprietary architecture, It is in their economic best interest to convince the world that it’s going to take forever for this to really open up. And then all of us who benefit from the opening up of RAN have various perspectives when that will happen.

I will tell you that at least in our opinion, it is inevitability. It is definitely a when, not an if question. There were many workloads in the IT datacentre that didn’t work great initially on virtualized infrastructure, but over time, the computer got faster the software got better. And now virtually every application in a data centre runs in a virtualized container. So I think your point there, where a lot of people [were] talking about it a long time ago – frankly predating Dell’s entry into Telecom – for those people, it’s happening much slower than they predicted.

From our perspective, there are absolutely people pushing it and kicking the tires. But it’s going to be a progression – today’s RAN through to virtualized RAN to Open RAN the spec. There were a lot of people that thought we would just flip it quick and go from yesterday to a future where everything was open, and you can pick a radio from any vendor you want, RAN software from any vendor you want and run it on any generic or Intel architecture servers, and that’s just not the way these things work.

It’s interesting what you said there about the telecoms industry moving more towards your area of specialisation. Do you think Open RAN is perhaps the first of many ways the telco sector and the wider tech sector are going to start merging?

There’s really a fusion taking place –  [telco engineers] are not unfamiliar with cloud architecture, they see it in their IT department. They know what an Amazon looks like, they know what a Microsoft Azure and a Google Cloud looks like. And they know what they’re doing in their own IT department which is embracing cloud architecture. And so they think ‘why is our network architecture so different than what seems to be the dominant way to deliver IT services?’ And isn’t communication just another information technology service? Particularly in the IoT, data, edge era – whatever edge means to everyone.

The notion that we’re going to be delivering more and more interesting applications from further out in the telecom network means that that fusion you mentioned has to happen. My background is in the information technology industry, not the telecom sector – and this is not a new trend. I think a lot of the people who were born and raised in telecoms say ‘oh my goodness, can you see what’s happening, it must be the first time in history!’ No. The RAN is a very, very specialised high performance workload, but there are others.

HPC, data analytics, artificial intelligence… there are other super real time, high CPU power consuming applications and what will happen is the architectures will evolve to enable those apps. I read a couple of your articles recently and to the point you make, a lot a lot of the analyst predictions of how much economy there will be around 5G rest on things like enterprise applications which will be delivered from within that network.

It’s a complicated relationship – many operators are putting over lots of their own systems to the hyperscalers, and also lobbying for the big tech firms that run services like that to start helping to pay for the network infrastructure, because they argue they’re the ones actually making money out of it. It seems as they’re getting closer together, there’s more and more adversity in some respects.

I think 5G is seen as the enterprise ‘G’, and potentially the last bastion of how a network operator can dip into the profit pool created by their network and not have it taken away by other people. And there are lots of responses to it; one of them is the over the top players should fund building a network. This is kind of not how it works from a capitalist perspective, but if they want to try to do that…

Most of our customers are all very focused on ‘how do we build value in our network that allows us to tap into these new profit pools and not have what happened with Netflix and other things happen again’. And most of them have concluded they need to embrace cloud architecture, in the sense that they need to not only resell certain capabilities and applications from others on their network, but they need to develop some of their own and sell them as well. I mentioned earlier the computer industry moves in cycles, and it’s like a pendulum swinging between centralised architectures and decentralised architectures.

Way back at the emergence of the mainframe, it was a very centralised architecture. All the data and applications ran in one place, and dumb terminal tethered to it. And then we moved into an era of decentralised architecture with client/server/internet. And all of the devices we now use are very smart, but they also have connections. The public cloud era has been kind of a reversion to large data centres – not mainframe computers – but large data centres. ‘Give us your applications and your data goes on here.’ And now we’re beginning to see the pendulum play once again to a distributed architecture called edge, where applications will run on place other than a company’s data centre or a public cloud.

And in that regard, the world’s network operators have the most valuable real estate in the world. At the end of the day, they own an awful lot of edge. Some large enterprises will likely build their own edge infrastructure, but most of the world’s businesses are going to be renting it from telcos. If they have some edge computing they need to do for some applications, there’ll be deciding ‘is it better to write this application and demand that the server is on my premises, or is my connectivity with my network operators sufficient that I can simply rent it from them?’

As that emerges, most of the world’s network operators hope to play a big role in that. And it’s just another example of the collision of IT and network communication.

Read the other half of our interview with Dell here, in which Hoffman talks about chip shortages, AI, the supply chain crisis, and how the pandemic has changed working culture for good.

Get the latest news straight to your inbox. Register for the Telecoms.com newsletter here.