I’ve noted in past blogs on MWC that the cloud providers were front and center at the show, and also that public cloud providers had their sights on the telco space in general, and 5G in particular, as a conduit for future revenues. Amazon sent a copy of their MWC presentation, and it’s a good way to analyze what cloud providers are saying and thinking with regard to their future relationship with telecom. The key slide in their presentation shows three layers of AWS support for telcos, so we’ll look at them starting at the bottom.
The bottom layer supports telco OSS/BSS applications, and this is the layer that most telcos tell me is interesting to them. OSS/BSS is the “core business” application piece for the telecom vertical, and much of it is the same as, or easily related to, core business applications for the broader market. However, Amazon announced the AWS Telco Network Builder (TNB) at MWC, so let’s take a look at that.
The tagline for TNB is “Automate the deployment and management of your telecom networks”, and it’s an orchestration and management framework to deploy network applications using industry-standard TOSCA service definitions, and also with the NFV ISG specifications. Deployment options include AWS CloudFormation, EKS, and VPC. I think that TNB is essentially a TOSCA decoder that then links to AWS services and features, combined with virtual network function (VNF) deployment and management. As such, it is targeting network services rather than OSS/BSS, and it extends the “basic” telco support that vanilla AWS would offer.
Having support for VNF deployment is as good as VNFs are, of course, and that’s the big issue for TNB. While VNFs are indeed linked to standards (ETSI NFV ISG specifications), the NFV concept hasn’t exactly exploded on the scene. As I’ve noted in past blogs, operators have had very mixed views on the value of VNFs as elements in the service data plane, because of performance and latency issues. If VNFs are actually to be used by telcos, those telcos could look at hosting them on AWS using TNB, but public cloud VNF hosting is also a stretch. Given that VNFs themselves are limited, adding public cloud hosting raises the issues considerably.
There are places where TNB could make sense, primarily in control-plane elements of things like 5G. However, the “telco network” is now and will likely always be dominated by network devices and not VNFs. I think that, ironically, TNB success may hinge on the success of Nephio, which is a Google project that would enable Kubernetes management of network devices as well as virtual functions and other containerized applications/components. Will Amazon join that initiative, or expect telcos to integrate Nephio. If they don’t do either of these things, I think TNB will have limited impact.
Amazon’s “Telco Ecosystem” middle-layer piece is focusing on the partnerships they’ve established with vendors, integrators, etc. The next thing Amazon talked about at MWC was the top layer, where they introduced the Smart Home Platform, Integrated Private Wireless, and Network APIs.
The smart-home piece seems aimed at providing telcos a role in the smart home revolution. Amazon points out that consumers love the concept of a smart home but are often defeated by the technology requirements. Telcos could play a role in this, offering an ISV platform that would be device-agnostic and centrally managed. I think this could be a smart play, but more for smart buildings than for smart homes, because I’m doubtful that consumers would pay for the integration given that they can get apps and hosting from the device vendors, including Amazon with its Ring series.
There is an extension of the concept for businesses, though, and that seems aimed at supporting IoT, AI/ML, and other emerging technologies through the cloud. Here there’s already cloud acceptance, and Amazon is working to frame their cloud services to facilitate development by ISVs in the key vertical areas. In some cases (connected vehicles, for example) they’re also addressing a related consumer-market opportunity.
Amazon, like others, is infatuated by private wireless, and there is a private wireless opportunity, but it’s one that takes some real effort to develop and is difficult to address at all except in a few verticals. My data says that Amazon’s customer base is only about half of Nokia’s base, and that its users are smaller as well. It’s interesting that they cite many of the same industries that Nokia does, including automotive/fleet/transportation and agriculture. Where Amazon is differentiated is in its ability to offer “managed private 5G” to SMBs as well as integrated operator/private 5G with partners in the CSP and MSP spaces. They work with integrators and radio vendors for the on-site deployments.
Integrated private wireless (IPW) seems to be Amazon’s main focus here, which I think is smart because it can support a broader set of missions and even a broader set of verticals than a pure private strategy. The partnering with CSPs and MSPs offers Amazon an entree into the telco space too, since most of the 5G elements and deployment and management tools are shared with Amazon’s telco service portfolio.
Not surprisingly, a lot of Amazon’s telecom story is dependent on third-party developers and hosted applications they build for verticals. Amazon has the largest developer community of the public cloud providers, and they also have a fairly well-articulated edge strategy that links with their core AWS services and APIs. That means that AWS can be bound tightly to on-prem edge elements (AWS Outposts).
Overall, Amazon is playing its cloud cards in the telco world, and all of their efforts are based on a fundamental truth, which is that in today’s market, “services” are international and telcos are constrained geographically. There is zero chance that telcos would deploy their own assets over their entire proposed service footprint because the cost of simply acquiring facilities would be crippling. The cloud is the obvious alternative, and it’s clear that Amazon is going to enhance the services AWS can offer (AWS Wavelength is an example) as well as exploit its base of developers and integrators.
Will that be enough, though? That’s the big question with Amazon’s telco strategy, and it’s a question Amazon probably can’t answer and may not even be able to influence. Amazon wins if operators, or at least MSPs, decide to build real higher-layer services on top of connectivity, services that may not require tight coupling with what’s below but could benefit from it. Amazon loses if operators stay with connection services, because truth be told there’s not much to be gained by the limited VNF hosting opportunities that don’t involve the data plane. For the data plane, white boxes are the cost-cutting (and feature-enhancing) answer.
That doesn’t mean that the Amazon telco initiatives aren’t important. Startups have been AWS bread and butter, and they’re in decline at least for now. Microsoft is doing better in the enterprise space, and will likely continue to do that. Google is still groping for a differentiating strategy, and for literally decades, operators have been more comfortable with Google as a partner than Amazon. If Google finds their feet, they’re more likely to take share from AWS than from Microsoft. But…and it’s a big “but”…if Amazon can be a bit less pedestrian in their thinking about the telcos, and work a bit harder to create a real opportunity for a JV, they could marginalize Google, and they’d then likely gain more from that than rival Microsoft. Think about it, Amazon.