One of the question marks over cloud-based core and RAN systems is how far they will be dominated, on the cloud infrastructure level, by established players like IBM Red Hat and VMware, and how far they will move to new platforms run by webscalers or new entrants. Among the cloud majors, Google Cloud has lagged behind Amazon AWS and Microsoft Azure in market share, but in the past year has made a very targeted bid for the high growth that is expected in the telco cloud sector, on the back of 5G and edge computing.
Its killer weapon is Anthos for Telecom, which brings its Kubernetes-based Anthos cloud application management platform to the network edge, and optimizes it to support telecoms-centric applications such as AI-enabled customer service or personalization systems.
Its latest step forward is a new collaboration with Intel, to develop cloud-native 5G technologies and services for operators, enterprises and even vendors, based on technology components from both sides.
Intel is supplying its FlexRAN reference platform for 5G virtualized RAN, its cloud-native Open Network Edge Service Software (OpenNESS) deployment model, its Data Plane Development Kit (DPDK), and hardware infrastructure based on Intel Xeon processors. These will be integrated with Google’s cloud platform and Anthos for Telecom, and Intel will add reference architecture and solutions to accelerate 5G vRAN with Anthos.
This is another feather in the cap for Intel, which has a growing list of partners that are developing and deploying 5G vRAN solutions based on FlexRAN. This reference platform has won Intel a place in many 5G trials and helped it gain an entrenched position even before any processor architecture has been deployed at scale for vRAN (there are plenty of challengers, most directly Marvell, but with more orthogonal plays from Nvidia, Qualcomm and others).
For Google, it draws on some of the advantages it claims over AWS and Azure in the 5G space, particularly a highly edge-centric approach. “We are linking Anthos and OpenNESS together so that the telecom service providers have an open platform that can deliver cloud capabilities across multiple environments in the Google Cloud, in the telecom cloud, or a telecom edge location, or all the way to an enterprise edge location,” Shailesh Shukla, general manager of networking at Google Cloud, said.
This should streamline complex processes of integrating, validating and testing network functions and use cases, as operators start to disaggregate hardware from software and adopt multivendor cloud-native functions (CNFs).
Google and Intel plan to establish multiple labs, and the first is already operational in New Mexico, USA. They said they will also open a lab dedicated to validating CNFs with multiple vendors, and creating blueprints for enterprise-specific 5G edge applications for sectors such as manufacturing and retail. The aim is to pre-integrated and tested use cases that can be deployed by operators or enterprises in a largely pre-validated way.
Shukla told SDxCentral: “I think fundamentally operators want this type of an arrangement. … The number of use cases are myriad, and which one is going to be the likely killer use case by vertical is not something that is known today. This is where the platform approach is important because it’s not as if you’re going to build top to bottom for one use case.”
The collaboration is clearly focused initially on 5G enterprise applications, and it could help Google to differentiate itself against Microsoft’s Affirmed/Metaswitch strategy in supporting 5G services for industries and their service providers. Eventually, it may lead to Google Cloud-based implementations of full-scale 5G vRANs. For now, the performance requirements and business models are uncertain, but it is not too fanciful to imagine future greenfield or Tier 2/3 operators turning to one of the webscalers for a hosted approach to 5G.
A year ago, Google Cloud unveiled Global Mobile Edge Cloud for 5G and announced a collaboration with AT&T on 5G and edge. It has added customers such as Telefónica, Orange, Telecom Italia and most recently Telus of Canada to its list.
The Global Mobile Edge Cloud platform aims to provide a more open and collaborative approach to the telco edge than rival offerings like AWS Wavelengths and Azure Edge Zones (though Telefónica has also signed up for Azure Edge Zones). Google’s model and infrastructure are less centralized and more network-centric than those of its rivals and that may enable it to form more multi-faceted collaborations with operators.
Under its edge banner, Google says it aims to deliver:
- a portfolio and marketplace of 5G solutions built jointly with telcos
- an open cloud platform for developing these network-centric applications, in which Anthos is the cornerstone
- a global distributed edge for optimally deploying these solutions, which will be developed rapidly by “lighting up thousands of edge locations that are already deployed in these telecom networks” rather than building all the sites itself
“We’re basically giving you compute power on our edge, where previously it was only for Google use, through the Anthos platform,” said Eyal Manor, VP of engineering for Anthos, last year. “The edge is very powerful, and I think we will now see substantially more innovation happening for applications that are latency-sensitive. We’ve been investing in edge compute and edge networking for a long time in Google over the years for the internal services. And we think it’s a fairly unique capability now to open it up for third party customers.”