This week, APPLEDORE RESEARCH GROUP attended NOKIA’s Global Analyst Event, the first since the merger of Alcatel-Lucent into the NOKIA family. Approximately 40 analysts arrived outside Helsinki, Finland for two days of presentations, demos, 1:1 briefings, deep dives with experts, and of course, beautiful Finnish summer.
Kicked off by NOKIA’s CEO, Rajeev Suri, this event began with a very positive commentary on the synergies provided by NOKIA’s and ALU’s complementary product strengths, technologies, and customer bases, and pointed to strong margin performance and quality metrics – which are often difficult to achieve during the process of integrating two large firms. Suri went on to paint a broad picture of the industry’s direction, and NOKIA’s position in it, describing a future of vastly increasing capacity demands, the need for low latency (~ 1 mS), order-of-magnitude increase in devices (many of them “things”) and falling ARPU (again, a result of simple “things”). NOKIA he says, is moving to deliver a network architecture that can meet these often conflicting demands based on a broad array of “5G” technologies, a highly distributed cloud architecture that pushes processing closer to the devices (latency, performance), and a “network operating system” that abstracts network functionality and exposes it to external applications. Fair enough. APPLEDORE RESEARCH GROUP commentary: This vision will demand very high performance from both the cloud infrastructure (VNFs), and complex and precise instantiation from the MANO and related management systems. It will also demand (and see more on this below) significant growth in small cells and simplified, low patency access – both fixed and mobile.
A succession of Suri’s lieutenants took each of these topics one level deeper, and I’ll try to give a very brief summary of each, paraphrasing for brevity and clarity:
- Samih Elhage, President of Mobile Networks, described the collection of 5G technologies that can deliver vastly more traffic, vastly lower latency, and vastly different demands – ranging from low cost and low traffic for sensors, to ultra low latency for VR and mobile control apps, to the continued march of video and other apps demanding of traffic. APPLEDORE RESEARCH GROUP commentary: Since we hit the Shannon limit with LTE, improvements demand MIMO, shaping and small cells. We also note that the Internet of Things is a highly diverse environment which demands that radio cost and performance both scale up (video, VR, control) and down (sensors, meters, etc.)
- Federico Guillen, President of Fixed, spoke of “fixing the unfixable”. One key take-ways were a) the economic value of moving to optical within enterprises – returning large amounts of office space previous used in copper closets. Another is the effective culmination of a process i have been watching, and (too early as it turns out) predicted ages ago – that fixed and mobile networks are converging, with more and more small cells effective transforming mobile networks into “fixed networks with wireless terminations” – which looks suspiciously like all our offices and homes with broadband and wifi – only it will be broadband to 5G plus wifi. APPLEDORE RESEARCH GROUP commentary: This vision demands that networks be managed holistically – no longer “mobile” and “fixed” – but as hybrid networks where access technology is varies, but the architecture is constant. It also suggests that mobility may no longer be tied to “wireless” nor “wireless” to mobility. IP provides a type of nomadic mobility; and devices are increasingly multi-interface. Increasingly the choice of copper/coax/fiber/radio will be one of construction practicality and economics – linking the “last 300 feet”.
- Sri Reddy, VP within IP & Optical Networks, spoke of the ever present “squeeze” between rising traffic, flat ARPU (let’s put aside changes due to IoT) and falling costs (attributable to advances in equipment and technology). His main focus was on simplification and costs reduction in the core, with a focus on simplification. In APPLEDORE RESEARCH GROUP’s opinion Sri is correct: a ubiquitous environment (whether its the IP network or NFV-I), is essential. The hyper scale cloud providers are proving this point with their virtually self-managing clouds. he went on to emphasize that one of the biggest hurdles to innovation and cost reduction is OSS – and THAT is a result of adding overlay silo after overlay silo, all of which become a spaghetti of systems integration and work-arounds. NOKIA’s solution is a simplified “smart fabric”, with the ability to tunnel multiple services automatically – ion effect WAN SDN. It was also interesting to hear about their work to make virtualized network functions — including packet intensive gateways — much more efficient and higher performing whether on generic X86 or NOKIA’s AirFrame. There is some independent testing that indicates that they are running as much as 4X compared to designs without their optimizations. “Four times exactly what?” is always the question of course — but the numbers were encouraging.
- Bhaskar Gorti, President of “Applications and Analytics” (which includes OSS, BSS, and MANO) provided and overview of their progress, the integration of NOKIA and ALU assets, and their areas of primary focus. Unsurprisingly, virtualization with the combination of Gorti’s CLoudBand assets along with SDN and AirFrame from “sister” units, was toward the top of his list. Gorti emphasized the commercial momentum of the suite, and the great synergy of NOKIA and ALU capabilities, making for a better whole. He also made a clear point that their goals (worthy ones in our opinion) are to 1) monetize, 2) Accelerate Innovation and 3) Improve the customer experience – and often to mix the three.
- Finally we heard form Marcus Weldon, President of Bell Labs and NOKIA’s in-house future seer. If you have not heard Marcus spin the future, you really should. He takes a true technologist’s perspective, and moves logically from fundamental trends and limits to define the inevitable goals we must strive for, and the limits we must live within. Weldon claims that these limits are 10-100x more capacity, correspondingly lower costs per bit, and 1mS latency — and all this means a highly distributed cloud, small cells, and intelligence likely < 25 km from many devices. This distributed cloud – “micro datacenters” in the webscale vernacular — favor telcos over the hyper scale cloud providers since only telcos have the 10s of 1000s of facilities pushed far out into the network. Think back to Suri’s view of the future, noted above. The story is a powerful one and is worth paying attention to.
We don’t really have space here to delve into the deep dive data that we received in day two, and in many 1:1 discussions; but come back often and I’m certain you will see some of the data worked into topical blogs or APPLEDORE RESEARCH GROUP research notes.
Thanks for reading!
Grant Lenahan