The Telco Cloud world shook today as AT&T and Microsoft made what could be one of the largest announcements of the decade, at least in directional importance. Appledore believe that this deal has far reaching importance because it clarifies which assets, skills, intellectual property and expertise are most fundamental to telecom. On the flip side, it clarifies which are less telecom specific, and best served by public cloud infrastructure giants. If it’s not telecom specific, those public cloud giants have engineering and deployment scale that cannot be matched by any single industry.
About a year ago we reported on Microsoft’s plans for “Azure for Operators”, in which they promised to bring the expertise necessary to operate high availability, highly scalable network workloads to the efficient shared resource model that is Azure. Not long after, Microsoft announced the acquisition of Affirmed Networks and MetaSwitch, claiming the intention of not only making these assets available “as a Service” but deeply learning from the specialized knowledge embedded in those acquisitions. We applauded both actions.
Today AT&T and Microsoft announced a deal in which AT&T will transfer significant assets into Azure for Operators, and Microsoft will operate the AT&T 5G cloud assets on the newly expanded Azure for Operators. Microsoft already has a very significant investment in its hybrid cloud Azure infrastructure (which we have covered deeply, here); AT&T will transfer software, expertise, existing infrastructure and even personnel with that elusive “network operational experience”. The combined Azure for Operators assets will (initially) operate AT&T’s 5G network from mobile core to MEC workloads.
There is some lack of clarity on what the initial workload complement will be, and how quickly the arrangement will move from mobile core toward networks supporting MEC, but both parties seem to concur that the plan is to operate AT&T’s end-to-end 5G infrastructure on the combined telco cloud solution. Regardless, according to Microsoft, all of AT&Ts 5G virtualized assets will be operated by Microsoft, even if they do not yet fully run entirely on Azure assets per se. Given the scale and scope of pre-existing facilities, this is the only practical path.
We have long argued that telecom ought not re-invent the wheel, and even more strongly, that early ‘virtualization” efforts were far behind the state of the art. During off-the-record discussions most telecom executives acknowledged that NFV was far from delivering on its promise – in fact it had in some ways take a step backwards, albeit along a learning path. In context this 3 year old survey of telecom’s cloud-native state-of-the-art seems strangely relevant and up to date.
The scope of this deal extends beyond AT&T as well, as, for maximum economic benefit, it must. What is being created is a telecom-savvy cloud native stack, that can be delivered (consistent with the Azure for Operators blueprint) in private or public versions, core or edge — “double hybrid”, if you will. The target is network workloads across service providers, bringing greater shared resources economies of scale; and support for any network workloads as well – not merely those selected by AT&T. After all, this is now a Microsoft product and when “aaS”, a Microsoft service to the industry.
Appledore have long argued that cloud demands thinking very differently. Both cloud economics and cloud resiliency depend on scale; on shared resources supporting many workloads; on active, hands-off management; and on standardized models and control loops – not bespoke solutions and logic. In one fell swoop, this announcement lays the groundwork for all of these transformations, and, in doing so, for a more competitive economic foundation for the industry (or at least those that get on board).
We applaud the direction and will report in more depth as soon as we have the additional insights to do so.
Grant