We normally pay limited attention to the musical chairs of executives within and between enterprises, but an exception can be made for the departure of VMWare’s general manager for telco and edge cloud, Shekar Ayyar, to establish his own start-up AdMY Technologies.
This is not so much a departure, more a strategic shift motivated by what VMWare, in keeping with other major enterprise IT technology vendors, has come to see as its biggest growth opportunity, the network edge. Especially the mobile network edge. The new company- not quite a spin-out because it is notionally independent of VMWare – is focused specifically on acquisitions in 5G and edge in areas complementary to VMWare’s own strategy.
What this signifies is that VMWare has acknowledged that its own monolithic approach to innovation around virtualization and edge computing based around organic growth and development is unsuited to the pioneer spirit of 5G-related edge compute. As Ayyar himself stated: “This acquisition vehicle structure will greatly enhance the ability to identify, target and invest in candidates within the telco, 5G and edge industry.”
VMWare, for its part, pointed out that Ayyar would continue serving as special advisor to its interim CEO Zane Rowe through the second quarter of 2021 and that he would also, on a longer term basis, seek opportunities to work through acquisitions with the company’s telco and edge ecosystem.
It is certainly true that competition among enterprise vendors around the network edge is hotting up. Equally operators, as well as traditional cellular infrastructure and technology vendors, know that they must embrace the established expertise in virtualization of VMWare, as well as the other big enterprise IT players like IBM, HPE and Dell.
The 5G camp may not entirely share the view promulgated by enterprise IT vendors that, while RAN disaggregation and ORAN are motivated by prospects of cost savings, future mobile innovation will come almost solely around the core and the edge. That is certainly a view we have heard expressed by Domenico Convertino, VP of product management for HPE’s Communication and Media Solutions business unit. It sounds more like a rallying call for enterprise IT, when the true picture is less binary – the cloud-native core, the edge and the disaggregated RAN are really inseparable when it comes to both innovation and cost savings.
It is true, though, that the underlying skills and technologies required for virtualization, and now containerization, of network functions have already been developed in the enterprise IT arena and just need adapting for the 5G edge.
Until 5G, the role of enterprise IT vendors in the cellular arena, apart from traditional corporate functions such as internal financials, was largely confined to subscriber and billing management. That is all changing under 5G as these vendors emerge to capture a substantial part of the ecosystem out as far as the RAN, especially at the edge. This is significant given the fundamental role edge compute is playing in meeting all three of the primary 5G use case groups, especially ultra-low latency.
Latency was at the forefront of IBM’s mobile edge developments, which focused on identifying which workloads and processes should best be executed locally to avoid delay. Part of this required ensuring that tasks not time critical are still centralized in order to dedicate edge resources to time critical processes, avoiding costly over provisioning there.
With this in mind, IBM came out with two complementary packages in May 2020, its Edge Application Manager and Telco Network Cloud Manager, developed in conjunction with its Red Hat subsidiary. Apart from latency, a key objective was to address the burden of managing increasingly large networks of edge nodes, with the Edge Application Manager incorporating machine learning-based analytics to manage up to 10,000 edge nodes simultaneously from a single administration point. This bore the fruits of IBM’s open source project Open Horizon, aiming to allow a single person to manage as big a network of edge devices as possible, with plans to go even higher.
Then the IBM Telco Network Cloud Manager runs on Red Hat’s OpenShift platform to orchestrate the virtual and container network functions that telcos will have to grapple with as they seek to inject NFV or containers into their infrastructures, to modernize for 5G.
HPE has been proceeding along rather similar lines with its CoreStack, unveiled in March 2020, although with greater emphasis on aligning with migration from 5G Non-Standalone (NSA) to 5G Standalone (SA), with its new 5G core, ready for network slicing. This HPE platform is served pre-integrated on carrier-grade infrastructure with embedded automation designed again to handle large numbers of edge nodes.
HPE has also placed store on extending the case for enterprise IT in 5G from one merely of delivering benefits of cloud-based virtualization already proven in the general data center to address emerging challenges operators will have. One such challenge is to ensure that the higher integration costs associated with virtualization and disaggregation do not outweigh the savings achieved through use of commodity hardware and greater flexibility.
HPE’s 5G Lab in Colorado is dedicated to this task, opening its doors to both vendors and operators to collaborate over interoperability issues. However, it sounds as if the big infrastructure players are not welcome participants given HPE’s positioning of the lab as a vehicle for operators to escape their dependence on the big three of Huawei, Ericsson and Nokia.
Indeed, the growing presence of enterprise IT players can be seen as a threat to those established players with the emphasis on virtualization and also efforts to woo their own traditional customers with packages for private networks, avoiding need for public spectrum. This trend was evident in VMware’s collaboration, announced in February 2021, with Dell and South Korea’s SK Telecom for a package called OneBox MEC, described as “a single box approach that provides enterprises with an integrated, private 5G and edge computing platform”.
Incorporating Dell EMC’s PowerEdge XE2420 server, VMware’s Telco Cloud Platform, and SK Telecom’s 5GX MEC Platform, this is targeted at enterprises wanting to deploy MEC inside their own premises, with SK Telcom identifying smart factories and hospitals as potential customer verticals.
The alternative approach to serving such use cases would be to offer edge compute as a service distributed within a public 5G network, without onsite deployment. That could equally meet latency targets but the onsite approach is favored by some enterprises for security reasons. Either way, these three partners believe that enterprises will be interested in such MEC packages for latency and reliability. They cite research they commissioned themselves from STL Partners finding that 40% of enterprises reported issues with latency of their current networking deployments, while 45% expressed reliability concerns.