VNFs on public cloud: Opportunity, not threat
There have been very few significant deployments of VNFs on the public cloud, even by leading telco advocates of virtualisation and cloud-native. STL Partners recently outlined extent to which telco network functions (VNFs and CNFs) are being deployed on public cloud, and particularly that of the leading hyperscalers. See AWS, Azure and Google.
STL Partners believes the continued separation of the network and IT domains in the thinking and practice of leading telcos (i.e. telco NFs kept separate from public cloud) reflects the persistence of a legacy concept of the telco, and an unwillingness to embrace the full logic and benefits of the cloud. In particular, the very notion of a ‘private telco cloud’ seems anachronistic: it harks back to when the network did truly represent a distinct technology domain, i.e. when it was based on dedicated, function-specific equipment. In the wake of virtualisation, and now in the era of cloud-native, the network itself (at least, the intelligent systems that run it) has become pure software and computing. And this is software, moreover, that is deliberately designed to be portable across multiple, hybrid cloud environments.
By resolutely holding on to ‘ownership’ of their telco cloud and CNFs, telcos are in effect trying to artificially perpetuate a separation of the network from the cloud that is no longer valid, as the network has become cloud. This does not, however, mean that there are not other valid competitive, regulatory or indeed security reasons why it might be desirable to maintain some degree of separation between the telco and hyperscale cloud.
Indeed, behind the concern that the ‘core’ connectivity business of telcos could be swallowed up by hyperscalers – if telcos became dependent on them to run their networks – lies perhaps a more grave concern: that telcos would then lose the logical network as a control point allowing them to dictate the terms on which hyperscalers gained access to their physical infrastructure as part of new edge compute use cases.
The nightmare scenario would then be: hyperscalers dominating both the cloud and the network – the edge compute, the application software (via partners) and the delivery mechanism (via the network) – with telcos being relegated merely to construction, operation and maintenance of physical infrastructure at the behest of the hyperscalers, and their application and connectivity partners. Almost literally just subcontractors digging holes in the ground, and wiring up cables, antennas and access points.
Over the past two to three years, many leading telcos have concluded co-operation agreements with the three largest hyperscalers that grant them access to the 5G edge to offer it as just such an edge compute platform to their application partners: Network-as-a-Cloud Platform (NaaCP).
Telco and hyperscaler co-operation agreements 2019 – January 2022
Source: STL Partners
NaaCP will drive more VNF deployments on public cloud, and opportunities for telcos as outlined in STL’s Partners report VNFs on public cloud: Opportunity, not threat.