While our natural desire to see and experience new places has been so abruptly curtailed, our thirst for knowledge is as strong as ever.
This week’s Layer123 World Congress attracted a sizeable online crowd to catch up on the latest developments in, well, just about everything at the IT/Telecom interface: SDN, NFV, Orchestration, Automation, Edge, 5G (of course).
There was competition from week 2 (or is it 3?) of the TM Forum’s Digital Transformation World, the Broadband World Forum, and a new Open RAN Forum to name just a few.
So if you joined one of those others, what did you miss at L123WC? Here’s a rundown of Layer123, from A to Z:
A – Automation. While fast becoming a term that is meaningless without some qualification, greater automation is a stated goal for most virtualization programs presented at Layer123. Orange’s Emmanuel Bidet reminded the audience that “automation does not come for free”, and that a route must be navigated between its benefits and the extra complexity that it can introduce. Liberty Global sidestepped the automation vs jobs elephant by framing automation as necessary for coping with a massively increased rate of change on networks. (By contrast, Deutsche Telekom offered a straight “fewer people” benefit on plans to disaggregate the access network).
B – Business Problem. Layer123’s roots are as a technical conference, but it was still surprising to see very few presenters call out specific business problems and target outcomes. That’s not to say there aren’t benefits. It’s just that without executive ownership (mere sponsorship isn’t enough) against a top-level, measurable business KPI, stalling is only a reorg away. In a panel, BT Global’s CTO Colin Bannon cited some 5G trials at UK ports but these still seemed at an experimental stage, more than market-defining.
C – Culture Change. Not for the first time, almost every presenter referenced the need for a change in telecom culture, as part of the inexorable move to more software-defined networks. This isn’t just a technical culture change. Liberty Global’s Director of Strategy Roadmap & Innovation, Chris Aspell flagged the challenge of getting a conversative Finance department on board. Inmanta’s Stefan Walraven highlighted how automation programs risk failure by not challenging basic thinking (with “brutal testing” echoing DT’s 2017 “brutal automation” speech).
D – Disaggregation. Opening up the black boxes in every part of the network. Not only mobile RAN but fixed access too, as Deutsche Telekom explained. The prospect of working with component parts, based on standardised interfaces, and banishing the dreaded “vendor lock-in” is just too appealing for telcos to ignore. Few concerns about the “free puppy” potential, but perhaps that reflects greater confidence from a year or two of PoCs and the fruits of large-scale collaborative effort like TIP and O-RAN.
E – Enterprise. The demands and expectations of enterprises loom large in SDN and NFV. In fact, few speakers referenced consumers at all. It’s clear that returns for telcos from structural software-ization are predicated on enterprises having ever-changing and ever-growing requirements.
F – Five G. There can be no “5G” strategy without a strong and comprehensive software-ization strategy. The good news is that 4G provided a lot of the groundwork. However, the way in which 5G radio allows more sophisticated use of spectrum points directly to the need for greater sophistication at the edge of the network.
G – Generation. As in 2G/3G/4G/5G. The more enlightened realise that virtualization/software-isation is not exclusive to any one mobile generation (as Mavenir’s packet.io acquisition also illustrates). And that (as mentioned in many other conferences) customers don’t really care. We will inevitably continue to evolve the access network, but what matters is what customers are able to achieve when the full power of the cloud, intelligence and applications can truly be accessed from anywhere.
H – Hyperscalers. A strange word that telco uses with an insinuation of sinister threat, while everyone else just gets on with it.
I – Industrial applications. A view is dawning that the primary commercial rationale for telcos to justify “5G” has to come from better-connected industry. Manufacturing, logistics, finance. Healthcare was less in evidence (and mercifully few references to self-driving cars) this year’s event.
J – Just do it. Multiple presentations advised against waiting until standards are fully finalised. Or the scope of programs fully defined. Or every business benefit itemised out. The future won’t be so fully mapped out that we can simply step into it. It’s only by getting on with it that the challenges will become apparent and ultimately overcome. (In a separate J, and a telecom conference first, Telefonica’s Diego Lopez invoked the Argentinian writer Jorge Luis Borges’ thought-provoking, metaphysical work in considering the boundaries and granularity of digital twins.)
K – Killer App for Edge: There isn’t one. At least, that was a takeaway from a discussion with STC, Spirent and others. Plenty of potential applications, and infrastructure flexible enough to deliver them will be the point, more than a single killer app.
L – Lean. As in “Lean NFV”, a stripped back architecture for NFV, designed to overcome the practical problems of mixing and matching virtual functions from different vendors. Created by some of the people behind the creation of SDN its pedigree is solid, even if it highlights a certain lack of thirst from vendors to drink the whole bottle of NFV Kool-aid. Colt’s Mirko Voltolini is currently piloting. More on Lean NFV here.
M – Multi-tenancy. For the CSPs in the (virtual) room, there was repeated encouragement to create edge clouds that provide the location for virtual functions for many different enterprises.
N – Network Orchestration. Kaushal Raman and Ashwin Menon gave a clear account of Verizon’s network automation journey. The practical challenge of having to upgrade thousands of enterprise sites for a single customer (such as a retailer) was one of the few real-world business problems that got airtime.
O – Open (in general) and Open RAN in particular. Open source software continues to be a major pillar within operator virtualization initiatives. The Open Networking Foundation showcased notable successes from its work:
P – People. Wouldn’t everything be so much simpler with fewer people in the mix? But the processes of change still need people to drive them, and where initiatives are falling behind, it largely seems to be people factors (or “lack of people” factors, ironically) that seem to be the issue.
Q – Quick wins. It’s not feeling like SDN/NFV is the place to look for those. Generational change, fundamental structural shifts come with the territory. Software-ization means multiple groups have got to move together, and that is going to continue to take time.
R – Rakuten. Arguably the best current example of a disruptive player in the industry. Enough said.
S – Systems Integration. Open’s dirty little secret that operators seem to be struggling to keep hidden: with all of these opened-up components, someone is going to have to do the plumbing… Security was also a key topic, with the shattering of conventional network boundaries posing new challenges. (Security in the Network Cloud is covered in new research from Appledore). Telecom Argentina illustrated the point nicely:
(Figure from Telecom Argentina’s Miguel Angel Masache Ojeda)
T – Trillions. Notwithstanding downgraded expectations for the commercial value of SDN/NFV as technologies, blue-chip presenters such as Intel are happy to attach figures in the Trillions to the economic value of the applications that “5G” makes possible. ONF’s Assem Parikh called fixed access network disaggregation telecom’s “trillion-dollar opportunity”. (This week also saw Nokia’s prediction of an $8 Trillion economic payoff from 5G-enabled industries…)
U – Un-telco. Actually, I believe PCCW Global’s Sharar Steiff referred to PCCW Global as an “ex-telco”, but many presenters reinforced the same viewpoint: telco’s future lies in them being a lot less like a telco.
V – Verticals. By now, enough work has been done to suggest that the technical concepts work. Now the question is moving to one of application – and to end-customer benefit. Telco cannot view verticals simply as a way of allocating sales resource to pitch standard products. A deep understanding of the business issues that particular industry verticals face has to be central to how telcos engage. See also I.
W – Wifi. How nice was it to be at a conference and not have to struggle with the venue-provided, vendor-sponsored, rubbish wifi? We’ll take our Covid-19 benefits wherever we can. On another note, speakers referenced the idea that telcos should offer hybrid wifi plus 5G using small cells for in-building coverage.
X – multipliers.. as in “10 x”. There was much less of this in evidence than in previous years. Benefits of virtualization seem to have been wound back a notch. Overall, the talk was of double digit % improvements in processes for configuring and managing networks.
Y – YANG. Few people now see the irony that the “YA” in YANG stands for “Yet Another”, but it seems to be broadly accepted as the basis for device and service modelling.
Z – Zero Touch. Layer 123 has been the backdrop for announcements on zero-touch initiatives in the past, even spawning its own sub-event. Much less of a key theme this year, while people work out whether it is a goal in itself, or just a special case of Automation. See A.
As a conference, the Layer123 World Congress has a vast scope to cove, across multiple areas each of which continues to evolve. As such, it reflects the huge range of challenges and contexts that operators and vendors are facing across the software-defined landscape.
The seismic impact of Covid-19 has massively upped the uncertainty factor, so perhaps simply building for the unexpected is as good an objective as any.