An interview with TM Forum, in two parts, about where telcos are up to with network autonomy, and the work to scale point solutions into vendor-agnostic deployments across entire network operations.
In sum – what to know:
Manual to partial – The telecoms industry is somewhere between Level 1 and Level 2 for network autonomy, says TM Forum, which holds the blueprints and the rubber stamp for how to get to Level 4.
Runners and riders – China Mobile leads the pack with autonomous fault management deployed at scale on Huawei and ZTE kit in certain regions; Telefónica and Orange are leading the European charge.
Blueprints for scale – TM Forum is developing broad vendor-agnostic blueprints for operators to deploy Level 4 across their operations, extending its ODA principles from OSS/BSS transformation to network automation.
Who’s the best at this? “China Mobile.” What about in Europe? “Telefónica and Orange.” And where is the industry, at large? “Between about 1.5 and 2.1.” So how does it get to ‘level four’, then – representing a kind of peer-reviewed vanguard for realistic near-term ‘high’-end AI-led network autonomy, better than partial level-two or conditional level-three autonomy? By stitching and stacking a series of use-case “scenarios” across an entire network operation, responds George Glass, chief technology officer at TM Forum. So how do you do that? Well, by “abstracting” vendor idiosyncrasies into a standardized blueprint, he says – so that any operator can apply it in any vendor configuration.
Something like that, anyway. But some history, first – as told to RCR Wireless by Glass. Because next month, more industry types than ever will descend on Digital Transformation World (DTW Ignite) in Copenhagen next month (June 23-25) to (mostly) discuss just one thing: autonomous networks. Which is a success story, no question, for an industry organisation best known for its important, but deeply unsexy, under-the-hood technical works (SID, eTOM, TAM, and so on, plus it key ODA framework for modular cloud-native telco IT, and related work with APIs) in the yawning void of telecoms operational and business support systems (OSS/BSS). But all of this is coming together, and it matters.
Because autonomous networks, increasingly using AI, are about the only story in telco town right now, and TM Forum holds the rubber stamp – with the “KCIs, KEIs, and KBIs” to measure how autonomous they really are. “We’re not talking about two percent or five percent,” he says. “We have seen 30 percent improvements in the mean-time to repair (MTTR), and 15 percent reductions in tickets – because they are not automating existing processes, but building autonomous capabilities.” That last part is important, of course. He says, later: “That is where we’re trying to join the dots. If you leave your legacy and just start putting agents on top, you will not get to level four.”
He adds: “You’ll get faster horses, but you won’t get tractors.” Telco transformation needs tractors – is the point. We will come back to it at the end, in part two. Let’s first replay the opening conversation in full.
Three telco missions
Glass is “babysitting” TM Forum’s autonomous network (AN) project for now, while it seeks a new “mission leader”, he says; he has broader control of its three-tier telco “missions”, all together: composable IT and ecosystems, autonomous networks, and AI-native operations and services – which effectively stack on top of each other to modernise the architecture (IT), automate operations (AN), and bring intelligence and decision-making on top (AI). “The autonomous network stuff is going really, really well,” he says. “It’s even surprising me – the momentum we’ve got on it.” That momentum, built over six years, has come to a head, it seems – like every telco wants an AN badge
Glass says: “Five or six years ago, five or six of us got together in a room at our Accelerate Europe event, talking about how to change the management and operation of telco networks using concepts such as intent, closed-loop controls, health-healing domains, evolving with AI, initially just to do anomaly detection, and ultimately to do automatic error correction and configuration of the network.
“We started with that premise, and we defined AN levels from zero to five, and kept that definition. We now have the capability for telcos to measure their AN levels, and also get verified and certified – to say, yes, you have indeed achieved 3.8 or 4.1. And that has put us on the map.”
Which explains, if it needs explaining, why level-whatever autonomy is being quoted in press releases and keynote presentations everywhere – including at “networking events that TM Forum wouldn’t have previously had a presence at”. But TM Forum, suddenly like the favourite trade-fair pitstop for every telco CTO (and CMO), is going further. Glass says: “It’s not just about defining and measuring AN levels; what we are trying to do is to show our members how to implement AN levels.” Right; but let’s back up first – because the rush of telco marketing proclamations about high-level AN certifications and ambitions gets in the way. Where is the industry actually up to, zero to five?
“Probably somewhere between about 1.5 and 2.1, realistically. It has got some automation of existing processes – is probably the best way to describe it… They typically come out at the high ones, low twos [in certification].” Okay, and who’s the very best telco in the world at this? “At the moment, from the evidence we see, China Mobile is leading the pack. It is rolling out IP fault management, broadband complaint handling, RAN fault management, core network fault management – and it is doing it with 300 million subscribers in Guangdong, 330 million subscribers in Xinjiang. Which just puts it at a level above everybody else.”
The industry has taken notice of China Mobile’s results, apparently – which represent an advance even on Glass’ initial samples, as quoted above: a 50 percent reduction in IP MTTR fixes in Guangdong; an 87 percent reduction in core-network MTTR fixes in Xinjiang; a 65 percent reduction in complaint handling, also related to the core network, also in Xinjiang. “Those are big numbers, which got the industry very interested in what we were doing,” he says.
Wall-to-wall autonomy
So what about in Europe? Or America? Glass responds: “Telefónica and Orange are making lots of noise, following our patterns. Telefónica has a number of certifications. Orange will have some too; it has done lots with fault management, lots on its network change. It is leading the way. Jio is interesting to watch; we’re doing certifications with Jio, to be announced at DTW. So it has made good inroads. The Americans are a little behind. How do I say it in a non-geopolitical way? The Chinese equipment vendors are ahead in their thinking and capabilities; the European and US vendors and operators are playing catch up.” That said, these others are making up ground, it seems.
He responds: “Mavenir is doing some cool stuff – a bit on the edge, sort of future. Definitely, Nokia and Ericsson are performing very well. At DTW, Ericsson will demonstrate both wall-to-wall fault management and dynamic 5G slice management – which is an interesting concept where you can turn the bandwidth up and down during a session by adding and removing slices in real time. Which it told us at the start of the year wasn’t possible, and we’ve now worked out a way to do it. Ericsson is… catching up very rapidly.” A couple of points to clarify: firstly, and despite all the telco press missives about level-four achievements, including their certifications, that 1.5/2.1 average holds.
Even China Mobile has work to do, says Glass. “It has only covered eight or 10 scenarios; it’s not about to run out of work.” Equally, Telefónica’s pre-MWC announcement, say, that it has a dozen live level-four cases is a snap-judgement about a series of silo projects scattered across its different operating companies. Telefónica wants to reach level 3.75 everywhere in 2028, and level four in 2030 – which is probably a frame of reference, at least, for where China Mobile is up to, and its faster timeframes for similar advances. The discipline now is to string these ‘scenarios’ together, says Glass – to eventually link them across technologies, domains, entire footprints.
He explains: “The network is vast, the operator is vast – many technologies, many domains. We subdivided into a set of scenarios based on whether we had a technical solution package to implement the scenario, and which scenarios delivered the biggest return at level four. We identified 55 scenarios, initially, and 20 or 30 have been added. Of those 55, 22 were identified as high-value scenarios – things like fault management in the IP domain, fault management in the transport, energy efficiency in the RAN, complaint handling. At DTW last year, two or three members had implemented a set of scenarios. At Innovate Asia in November, I gave 37 certificates.”
He adds: “But remember, some members have a 3.5, 3.8, 4.1 [certificate], whatever, but it might be for the IP domain, only – and only for fault management. It’s not level four across the network; it’s for a particular scenario or technology within a domain. But something like fault management covers five or six different domains. At DTW, we will show wall-to-wall fault management where we join-up those solution packages – so you’re not just running fault management for your IP or transport, coordinated separately; you’re running it at level four across your organization.”
Vendor-agnostic blueprints
Back to China Mobile: its approach is different, says Glass, just because it is way ahead with its 5G rollout already.
“If you’ve already deployed your network, you aren’t going to invest in network planning – because the network’s built. So it will use the planning [solution] for its 6G network. Whereas others, just building or optimising 5G, will find that capability very useful.” China Mobile is important in this scheme, too, because it is pushing the whole industry, including TM Forum itself. “We’re trying to normalize that, and China Mobile wants it normalized.” He is responding to a prompt to clarify that China Mobile’s progress is not just down to better automation functions from Chinese kit makers. It’s not the point, he says; the TM Forum architecture levels the field, so every telco can raise its game.
“We’re trying to remove that barrier. China Mobile was pushing to standardize it. And if an operator the size of China Mobile pushes its vendors to normalize and standardize, they are [going to] respond positively.”
This article is continued here.
