Large NEPs that promote the adoption of their Service Management and Orchestration software are looking for a “new point of control” as network disaggregation threatens their business model, according to telco software giant Amdocs.
In a briefing with TMN at the Telecom Infra Project’s FYUZ event, Amdocs’ Neil Colman, Product Line Manager O-RAN Automation, and Joe Hogan, Networks CTO, said that operators must think very carefully about using their single Open RAN vendor as the provider of their SMO.
“Do you want to let NEPs sell you the SMO and orchestration? If you do that’s one of the [most] profound decisions you’ve just made, very early, that will come back to haunt you later. The whole tenet of Open RAN evaporates, or is substantially threatened, if you go down that road,” Hogan said.
“If you’ve gone Open RAN single vendor and they have the SMO layer, it’s going to slow you down. It’s like dragging an anvil,” Hogan said.
Amdocs’ view is that a single vendor RAN and SMO provider can use that relationship to slow down progress towards multi-vendor networks that give operators access to true best of breed technology.
Coleman said that’s because NEPs can decide and certify which apps can run on their ecosystem, and they can certify which vendors they will actually manage.
Hogan said that NEPs could also potentially slow things down by overloading the protocol with slight variances to core standards.
“I think when you’re as large as AT&T, you have their attention. If you were a European Tier One or Tier Two, that’s when you find it much more difficult to influence their roadmap.”
Hogan had said similar on a conference panel which was also attended by AT&T’s VP of radio access network (RAN) technology, Rob Soni and Vodafone’s Head of Open RAN, Paco Martin. AT&T is indeed following the RAN-SMO vendor approach, as it is using Ericsson’s EIAP as the SMO for its Ericsson-led RAN refresh. In Canada, Telus is using Samsung’s SMO for its rollout, in which Samsung will be the sole vRAN (CU-DU) provider, as well as providing the majority of its RUs.
AT&T’s Rob Soni speaks as Amdocs’ Joe Hogan (far right) looks on, during a panel at FYUZ24.
In fact, for AT&T the threat of a recalcitrant vendor slowing down progress is not so great, Hogan said. “I think when you’re as large as AT&T, you have their attention. If you were a European Tier One or Tier Two, that’s when you find it much more difficult to influence their roadmap.”
Speaking separately at a panel session, AT&T’s Soni said that what is important for AT&T is making sure it has “programmability” of the network, with visibility into all the layers of its network.
“When the RIC was born people initially thought they had the opportunity to create a layer that could effectively replace innovation in the RAN layer, and disruptors and even incumbents could close their roadmaps more quickly by innovating on that layer. The reality is, there’s still a lot of opportunity for differentiation in the RAN software layer itself,” Soni said.
That means creating an interface which covers the aspects of being service-aware, intent-based and AI-ready.
“Joining those three things together is important for us when we want to deploy networks at scale. So we continue to look for opportunities to leverage the relationship (with Ericsson) to create that programmability and visibility.”
Within that, AT&T still thinks there’s a need to create a layer above the RAN, because there is other information that it can use to enrich an rApp (“or even an xApp, potentially”, Soni added, somewhat surprisingly) to cause us to get a better experience in the network.
“So the advantage of having a strong RAN coupled with a strong programmability layer is that we actually get the best of both worlds.”
Whatever the strategy, it’s a live issue. Coleman said Amdocs was seeing a lot traction in the SMO and RIC space.
The language of Open RAN
“We’ve seen more RFPs over the last nine months on orchestration than at any other time we’ve been dealing with SON or other automations. Whether they’re going to Open RAN or not, operators are using this as a way of evaluating what they’re doing from a management layer. They’re using the language of Open RAN to define their requirements, even if they’re not doing Open RAN.”
Of course, for Amdocs, which Hogan revels in calling “the largest telecoms software provider in the world”, the threat switches the other way. Vendors successfully selling their SMO and RIC platforms as a part of Open RAN migration impinges on Amdocs’ own position in Service Orchestration, especially as those SMO platforms could in time be extended into other network domains.
Coleman said that this was a “nice healthy bit of competition. I think it’s a good thing in the industry that these tensions exist.” However, he pointed out, Amdocs is going to be brought in with the specific mandate to enforce third party capability.
Amdocs, then, sees itself as a neutral broker (“we need to be Switzerland”, in Hogan’s words), ensuring that an operator’s RIC and SMO platforms are truly open and committed to enabling third party ineroperability, as well as being able to take advantage of the work that Amdocs is doing to leverage Gen AI capabilities in network operations and network use cases. (Coleman said: “We’re building apps right now that feed into natural language interfaces within Gen AI so you can query the network via the app API. There’s some nice links you can do when you’ve got the OSS.”)
That can enable more horizontal automation capabilities, as well as providing a genuine commercial opportunity for third party app developers.
Coleman said, “In the past automation has failed because it has continually been funnelled down silos. Having an open environment for apps is probably the best thing you can do to get that automation.
“It’s also very hard to see how, if you’re a small app provider, you can make any money in this [NEP platform] environment. If you say we have a great app for automating neighbour lists you’ll be one of thirty companies with reasonably good algorithms. And then the price just goes right down. You need a variety of RICs and you need apps that can run across a bunch of RICs so you have a chance for individual app providers.”
Hogan said, “We’re projecting beyond the horizon but I hope when we get there there’s a much more horizontally scaled environment for the xApp and rApp companies to come in and show what they can do. And then the large companies like Amdocs hopefully have an SMO role for integration, test and packaging.”
