Telcos largely failed in the cloud market. As recently as two years ago, operators and webscale providers were at loggerheads, both fighting for the same revenues in cloud-based services – the former group emphasizing their connectivity and their network-based capabilities such as location awareness; the latter fighting with their sheer global scale from infrastructure to developer communities.
Now, the majority of telcos have backed away from the cloud market and are seeking deep partnerships with the webscalers, as seen in a series of announcements between various operators and AWS, Microsoft Azure and Google Cloud. Those operator assets, such as high quality connectivity and network-based services, are still in play, but designed to be complementary to the cloud giant’s own strengths.
The next step, for many operators, may be to abandon the actual network cloud – the building of their own cloud infrastructure to support their internal IT functions and even the virtualized RAN, core and transport networks themselves. AT&T, formerly one of the most aggressive operators about the cloud market, has not only offloaded most of its externally-facing data centers, but pledged a “public cloud-first” policy for its own internal IT functions. It is still building its own next generation network cloud to support its virtualized, software-defined networks, but partners like Microsoft are increasingly deeply embedded in that effort, and it seems more than possible that the telco will eventually move its networks to a third party cloud platform.
Will the same pattern be seen in the edge cloud? Operators, as we have covered extensively in Wireless Watch over the past year or two, are placing great hopes on a combination of the distributed cloud and 5G to deliver both internal efficiencies and new revenue streams. As a first step, operators argue, they can leverage their distributed locations – central offices and even cell sites – to house their own cloud infrastructure. This could be rented to enterprise customers or to webscale partners – since the latter do not have a highly distributed topology – and could also support the operator’s own localized instances of its 5G core and RAN. Those local ‘sub-nets’, combined with edge cloud processing and storage, could enable a wide range of new services from private managed networks for enterprises to enhanced AR/VR experiences for consumers.
Operators have more inherent advantages in the edge cloud than the centralized cloud, because of their well-placed locations, and because of the timing, as they start to deploy 5G, which will greatly improve the quality of many edge-centric services, especially those requiring high bandwidth or very low latency.
However, they also have the same disadvantages that forced most of them to back away from the central cloud market. These include:
- They have limited penetration, at any strategic level, of the enterprise markets which will generate the best revenue growth and margin in the edge cloud space
- They do not have a significant global developer base comparable to that of AWS, for instance, and although there are efforts – such as Ericsson’s EdgeGravity and Deutsche Telekom’s MobiledgeX – to develop a shared platform, the track record for operators to work together around common APIs is not good (remember the Wholesale Application Community?)
- Other efforts to create a unified telco approach to edge have been fragmented – there are too many initiatives, and few of them involve the webscalers. The new GSMA-led group may have a bigger impact because of the influence of that industry association, but on the other hand, GSMA tends to take a fairly conventional approach to its members’ business models.
- They are not global, and although they have strong experience of creating global reach at a low level through roaming, when it comes to services and customer relationships, there is little coordination, by contrast with the webscalers. This could be damaging both to economies of scale and opportunities with multinational enterprises.
Operators must work together globally to create robust developer platforms; to pool site and connectivity assets in order to achieve massive reach and scale; and to ensure that they control high value processes such as orchestration, slicing and software-defined networking (SDN) across all the network and cloud resources.
If they do not unify their efforts around these ambitious goals, the webscalers will no longer be talking about the relatively equal partnerships that we have seen between AWS and Verizon or Google and AT&T. Instead, their telco friends will be providing commoditized locations and connectivity, and the orchestration, developer platforms and enterprise contracts will remain where they are now, in the hands of the cloud giants.