An open and shut case: Part 2

• Open networking systems are gradually finding  their way into telecom networks
• But their introduction requires operational and strategic changes at the network operators that are embracing openness
• External partners are often needed to access the required new skills, but what does that mean for how telcos manage ‘Day 2’ operations?
• This was discussed during TelecomTV’s recent Open Telco Infra Summit

One of the big questions for telcos embracing open network architectures is – how are such networks planned, deployed and managed and who is best placed to perform those tasks? That was one of the key topic discussions during TelecomTV’s recent Open Telco Infra Summit, when the potential role of systems integrators was once again under scrutiny. 

In the first part of this two-part feature –  see An open and shut case: Part 1 – we looked at the trend towards openness, the slow pace of adoption, the skills challenge and the importance of open-source community engagement.

Now we’re looking at the ways in which operators can approach the potential of open networking system deployment if they don’t have the in-house talent and tools to do everything themselves.  

Systems integrators and third parties to the rescue?

This touches on a very live area of controversy still raging amongst and within telcos: The beneficial dynamics of openness might have been accepted, but does yielding influence over the infrastructure compromise a telco’s ability to perform to its maximum potential? If the answer is ‘yes’ then a reluctance to engage system integrators, outsourcing firms or other forms of technical burden-sharing is understandable.

But external help is clearly needed in many cases. Disaggregated networks are more difficult to monitor, manage and support, and Verizon had to completely re-tool to do it, noted Beth Cohen, Verizon Business SDN network product strategy specialist, and that process took several years and is still evolving, she added.

Even a giant like Verizon, with its vast depth of technical knowledge, needs outside help when it comes to cloud-native operations and right now, she said, work is still needed to tie in monitoring and observability to improve network application performance.

“This is what customers really care about and is an area that the open-source community can really help us with,” she said. “We need standard APIs to manage the enormous and complicated networks and really drive integration and orchestration.”

Wind River CTO Paul Miller agreed, pointing out that once systems are deployed, the challenges centre on ‘Day-2’ operations. Here, he said, you need an ability to upgrade and manage ‘configuration drifts’, ensure Kubernetes certificates are up to date, develop expertise on system debugging, and much more.

These capabilities are rooted in software and Miller thinks that’s naturally a domain for software specialists, many of whom have experience of cloud-native operations in the private network environment, and nearly all of whom don’t yet work within telcos.

So does this represent a massive opportunity for system integrators (SIs)? According to Vivek Chadha, SVP and global head of telco cloud at Rakuten Symphony, it does. “With the best intent in the world, the majority of CSPs won’t have the skill set and process mindset to embark on a fully cloud-native, disaggregated, completely DevOps-led rollout on technologies that their workforce is not yet familiar with,” he said.

“Some CSPs are investing heavily in building those skill sets in house, but the rest will react and respond to the need as the network evolves,” and that’s where the SI’s can ply their trade, he maintained.

But some are sceptical. Wind River’s Miller, while fully supporting the “disaggregation is hard” line, doesn’t think SIs can solve the problem for all telcos. “SIs play their role and they have a great deal of value to add for, say, the smaller scale operator that may not have the staffing and expertise to absorb this kind of technology in the way that a Tier 1 might be able to.”

An SI, he said, can provide network engineering and deployment services, but can’t address the fundamental issues that impact even the larger telcos. He explained that when service providers have problems with, say, a distributed Kubernetes infrastructure, they need to call the supplier of that infrastructure, and if it’s a problem with the application layer, it’s time to contact the application vendor.

Therefore what’s required, he said, is an industry-wide agreement to sort out these virtualisation and application layer integration issues well before the delivery of the software to the service provider: This can’t be solved by an SI because it’s not in the picture prior to software delivery. To really get the open technology adoptable, there has to be a system of reference platforms and mechanisms to ensure that the industry can work together to integrate prior to delivery to the customer.

 Telco structural change

Once the fully disaggregated and re-integrated network environment is up and running, what happens to the infrastructure providers and the vendor/telco relationship? Will we see the ultimate pick‘n’mix, with telcos buying best-of-breed components and casting aside the tried and not totally trusted monolithic vendors? 

Probably not, at least in the medium term.

Certainly there are identifiable gains to be made with an Open RAN deployment: According to its telco supporters, it offers a path to greater energy efficiency (which is so important these days) and might help shake up the concentrated telecoms vendor landscape and diversify infrastructure supply. Then there are a host of minor gains, such as improved fault management and visibility that comes from having an open system or even being able to gain security insights from a broader range of expert companies. 

But will it be seen to be worth all the disruption and upheaval when it finally comes into force, our specialist speakers were asked.  

“It will if it can prove its ability to change the business model, not just the technology model,” stated Appledore Research principal analyst Francis Haysom.

With Open RAN, said Haysom, there’s an opportunity for a ‘one throat to choke’ supplier – a new player with fully integrated offering keen to take responsibility and underwrite performance for an Open RAN – while the deployment of a RAN intelligent controller (RIC) opens the door to improved RAN performance and/or energy savings, he stated.

But overall the analyst claimed he saw little evidence that, by itself, Open RAN can move the innovation needle, as he elaborated in a conversation with TelecomTV’s Ray Le Maistre.

“At the beginning in 2020 we [Appledore] came to the conclusion that there wasn’t anything about opening up the fronthaul that would necessarily disadvantage the major players” and therefore expand the ecosystem of radio access network technology suppliers.  

Many of the new opportunities in Open RAN are some way in the future, he said, but the main thing to consider is that the whole infrastructure vendor market, including the major incumbent providers, will likely go ‘open’ once all the players adhere to Open RAN standards and interfaces and that at least 80% of Open RAN-enabled system business will continue to be won by the existing big vendors or by “new integrated vendors that come out of Open RAN, but present an integrated solution” and therefore offered a single throat to choke should anything go wrong or support or improvements were required.

In other words, one way or another, vendors could pull together an integrated offering made up of fully compliant Open RAN elements that could later be swapped out or augmented by the operator with enhanced replacements provided by the same or another compliant vendor. So it’s unlikely there will be a major changing of the guard, with the current dominant players squeezed out and a new breed of lithe and specialised disaggregated players coming in to replace them as had been hoped or feared. “We really only envisaged about 20% of the market being truly Open RAN multivendor… [constructed by] putting new things together,” Haysom suggested.

Following disaggregation and the ability of telcos to buy disaggregated components, it may take a couple of generations of technology swap-outs before anything startling can happen to the telco business models sitting on top. When that starts to happen, the telco ecosystem might begin to meet Haysom’s “business model change” criterion. “We’d heard a lot about the need for a more diverse supply chain, but when it comes down to it, the attraction of having ‘one throat to choke’ for the telco has kind of trumped supply diversity,” he said.

Haysom’s analysis is currently being borne out by telco procurement plans, with the most startling example being that of AT&T, which intends to spend $14bn during the next five years on the deployment of Open RAN-enabled technology supplied mainly by recent Open RAN convert Ericsson, with walk-on parts for Fujitsu, Dell Technologies, Intel and Corning. Significantly, this move effectively spells the end of Nokia as a major RAN technology supplier to AT&T (at least for now) – see Nokia licks its wounds over AT&T’s RAN rejection.

AT&T plans to have fully integrated Open RAN sites, comprising Ericsson systems and Fujitsu radios, installed next year and, by late 2026, the operator expects that 70% of its wireless network traffic will flow across open-capable platforms – a decision which provoked an immediate drop in rival Nokia’s share price.

This series of events and shifts seems to illustrate that Open RAN support (or promised future support) has become required table stakes, even for the incumbent vertically integrated players as their telco customers show an understandable inclination to keep their disaggregation options open. 

Openness and disaggregation clearly marches on, with Open RAN as its standard-bearer. It may take longer than first thought, and we may continue to see today’s incumbent players in prime positions, but disaggregation and supplier and system diversity will develop further over time – the ecosystem’s customers will simply demand it.

- Ian Scales, Managing Editor, TelecomTV