Google Distributed Cloud: Google Cloud Moves On-Premises

In early October, as part of the “Google Cloud Next ‘21” event, Google introduced Google Distributed Cloud with the tag line “In your Data Center, at the edge and in the cloud”. The announcement went on to describe how an enterprise can accelerate their cloud adoption through the use of Google Distributed Cloud and realize the benefits of cloud; namely, easier development, faster innovation, more efficient scaling, and a reduction in technology risk, despite the many workload restrictions that are placed on the enterprise today.

The Google Distributed Cloud announcement detailed a portfolio of solutions that extends the Google infrastructure and ecosystem to the edge and into customer data centers. It details multiple locations from where the Google Distributed Cloud can be run:         

• Google’s Network Edge with over 140 Google network edge locations globally.
• Operator Edge which is optimized for low latency and enabling the enterprise to take advantage of communication service provider offered 5G/LTE services.
• Customer Edge for customer owned edge sites or remote locations such as retail stores, branch offices, or factory floors.
• Customer Data Centers for customer owned data centers and colocation facilities where the restrictions discussed above apply.

The Google Distributed Cloud is built on Anthos, Google’s cloud agnostic, unified, container-centric platform for orchestrating and managing infrastructure and applications across multiple public and private locations. With the Google Distributed Cloud, Google is leaning heavily on Anthos, pushing it into the operator cloud as well as into customer sites to help enterprises to modernize their environments or migrate to cloud. Google wants to transform the enterprise data center and it is backing its own platforms to achieve this.

Google was referring to the restrictions placed on a multitude of enterprise workloads and datasets which means they have not been able to be migrated to the public cloud. These restrictions are placed on data movement and data processing by sovereignty laws or by industry governing bodies or legal compliance obligations. Other factors that have restricted workload migration to the cloud are proximity to downstream or upstream services, and specific low latency requirements needed when acting upon the workload results. There is also the inherent lack of conviction from industries to allow their Intellectual Property (IP) to move physically outside of their secured property, no matter what assurances the Cloud Service Providers (CSPs) give.

With this announcement, Google has provided the enterprise with an alternative to the traditional cloud environment when considering its modern workloads, but does it mean that the enterprise data center will start to become overly populated with Google appliances? Most enterprises will only have a subset of workloads that require such an on-premises solution, while the remaining workloads would still be candidates for running in the traditional cloud. What Google’s solution does is allow the enterprise to prepare all their workloads using a common framework and then to distribute them, flexibly, to the most appropriate compute solution.

None of the Google Distributed Cloud technologies are unique in the world of cloud computing. For example, the Operator Edge environment that Google designed to exploit 5G capabilities is very similar to the Amazon Web Services (AWS) Wavelength Zones and Microsoft’s Azure Edge with Carriers. Likewise, AWS and Microsoft have an equivalent to Google’s on-premises solution. So, is this announcement simply bringing Google up to the same level as the established CSPs?

Google is considered by many to be running a distant third in the CSPs race to transform the enterprise workload through cloud adoption. While Google may have gotten off to a slow start, taking its time to build a flexible product with an equally flexible supporting ecosystem could be a strategy that sees it gaining ground on its rivals. Anthos will run on any underlying cloud platform, meaning that an enterprise can use Anthos to manage its various on premises and cloud workloads and point them to any compute solution in an any location. It will be optimized to integrate fully with Google’s own solutions, but the enterprise will not be tied to them. This flexibility is exceptionally valuable to the enterprise as it provides assurances against vendor lock in, where the cost to an enterprise of switching suppliers becomes so great and so technically complex that it is not an option.

In sitting back and not being first to market with its cloud solutions, by taking the time to carefully consider what the enterprise actually needs, is Google’s timing spot on? Certainly, they will be well placed to catch the second wave and exploit a cloud savvy enterprise whose pessimism has been replaced with a demand for a powerful and flexible cloud solution that allows them to choose where their sensitive workloads run. Whether this flexibility alone will this be enough to tempt the enterprise away from the more established CSPs will remain to be seen, but with Google Distributed Cloud it is clear that Google is keen to compete in this space and is looking to partner with the enterprise to do so in whichever location the enterprise chooses.