Google Cloud yesterday expanded its infrastructure offerings with a suite of solutions that run on the edge of a network as well as the data centres of businesses that cannot easily adopt the company’s existing public cloud services.
The new Google Distributed Cloud essentially extends its hardware and software solutions to a wider base of customers, including telecom operators that need low latency at the edge of their 5G networks and government agencies that have to keep their data in-country.
Using the Anthos open-source cloud platform, the new solutions can be run on more than 140 Google network edge locations, a telco’s edge, a customer’s edge or customer data centres.
They are meant to be easy to manage and quick for applications to be deployed on, according to Google Cloud.
Getting onboard also means businesses can make use of the databases, machine learning, analytics and container management that the cloud platform offers.
At launch, Google Distributed Cloud can count Cisco, Dell, HPE and NetApp as partners that will support the service.
There are actually two versions of the offerings unveiled at the Google Cloud Next virtual event yesterday – Google Distributed Cloud Edge and Google Distributed Cloud Hosted.
The first one, as the name implies, is meant to run on the edge of networks. This is where low latency is critical and where a trip to the public cloud takes too long for businesses.
There are a number of uses here, for example, running computer vision or Google’s artificial intelligence (AI) tools at the edge. A factory that needs to remotely inspect its production floors, for example, can connect to this fully managed Google service.
Google Cloud is also targeting telcos that are rolling out 5G and other services to their customers. Using Google Distributed Cloud Edge, they can expect to run workloads on Intel and Nvidia technologies for their 5G and other use cases.
This actually builds on what Google has been doing with Ericsson and Nokia, in a bid to bring cloud-native 5G cores and radio access networks (RANs) to telcos and to develop their network edge to offer new services to their business customers.
The second version of Google’s new distributed offering allows organisations to run Google Cloud solutions on their own data centres.
Google Distributed Cloud Hosted is designed to run sensitive workloads, possibly from government agencies or those in regulated industries like banking that have strict data residency requirements.
Unlike running Google or third-party tools on the public cloud, customers here can get hold of a selection of similar valuable tools by running them on-premises.
No connection to Google Cloud is needed to manage the infrastructure, services, application programming interfaces (APIs) and tooling. Customers can set this up themselves or with the help of a Google partner, and a preview is out in the first half of 2022.
Whom will this appeal to? Mostly government agencies, such as those operating in Europe that require highest levels of digital sovereignty, which Google pointed to in its announcement.
Currently, it is working with T-Systems in Germany and OVHCloud in France to deliver Distributed Cloud Hosted.
While this is an important push ahead by Google, what matters ultimately is what tools these new offerings will be offering.
In particular, what among Google’s cloud-based tools, such as AI or machine learning, will be available for its on-premises version? How many can be run on customer data centres without phoning home to the public cloud?
That said, this is still a positive move forward, given how organisations are looking to gain flexibility in the way they develop digital infrastructure and new capabilities in a multi-cloud future.
The new portfolio of offerings will also help Google Cloud level up with cloud competitors that have been offering similar flexibility through their extended offerings of late.
Microsoft already offers Azure Stack, AWS has Outposts and Oracle provides Cloud@Customer, each promising to bring new capabilities into any setup that customers want.