Hybrid cloud + Google Anthos
For similar content delivered to your inbox once per week, follow me over at bitsofcloud.substack.com
As much as I have the tendency to think that the battle is over and cloud computing has completely replaced alternative models, the transition of enterprise workflows to the cloud is still in its very early stages. In 2020, it was estimated that only 20% of enterprise workloads were running in cloud data centers, with the bulk of the rest running in on-premises deployments. While the trend to move workloads to the cloud is accelerating, especially with tailwinds created by the covid-19 pandemic, by 2023 likely no more than 40% of workloads will be running in the cloud.¹
Many enterprises today run their applications in an architecture known as hybrid cloud.
A hybrid cloud — sometimes called a cloud hybrid — is a computing environment that combines an on-premises datacenter (also called a private cloud) with a public cloud, allowing data and applications to be shared between them.²
Private clouds are data centers owned by companies on which only their own applications are running and only their own data is stored. Public clouds are data centers owned and operated by cloud providers in which resources can be rented by users who want to avoid managing their own hardware. The “big three” public cloud providers in the US are Amazon, Microsoft and Google.
Why hybrid cloud?
Cloud computing offers massive benefits in flexibility and speed over private cloud (i.e. on-premises) deployments, but there are a number of legitimate reasons why on-premises deployments remain around and even offer some benefits over the cloud, given the current state of enterprise IT.
One reason why on-premises data centers remain a part of most large companies IT architecture is the time and investment that would be required to move all workloads to public clouds. While startups and newer companies can enjoy the advantages of cloud-only computing from their inception, most larger and more established companies existed long prior to the cloud revolution of the last 15 years and thus have already built their applications on top of private data centers. Moving apps to the cloud requires not only immediate engineering time to rewrite portions of applications to run in a different environment, but also re-training entire departments with the set of skills that public cloud deployments require.
Secondly, many argue that with huge workloads, running apps on-premises can actually reduce costs when compared with cloud computing. This point is widely debated, and it can be difficult to find a fair-minded view when cloud companies and on-prem infrastructure providers have vested interests in arguing for the state of the world as they understand it. But certainly in the short run, continuing to run apps on-prem when they are already deployed in that environment offers cost savings. Gartner has estimated that moving workloads to the cloud can increase IT costs in the first year of migration by over 60%.³
As long as on-prem deployments still offer some benefits over migrating everything to the cloud in one shot, companies will need help managing the complexities of IT infrastructure living in two entirely different environments (aka hybrid cloud).
Introducing Google Anthos
Anthos is a suite of products offered by Google Cloud Platform (GCP) that comprises a single set of tools companies can use to manage applications across multiple environments. Anthos can be used to save time in managing hybrid cloud deployments because it integrates with on-prem data centers and then provides a set of APIs matching those which are used to manage GCP cloud environments.
The diagram above was taken from a presentation given by Ankur Kotwal, a Google developer advocate, at a GCP conference in 2019. Of course, there’s a lot to unpack in the diagram, and explaining each of the components shown would be beyond the scope of this article. However, one useful example of how a single platform to manage hybrid infrastructure is useful is the Stackdriver component, which is a tool for application monitoring and log storage.
Stackdriver (which just in Q2 2020 was rebranded as one component in Google’s Cloud Operations package) was originally integrated with GCP via an acquisition in 2014 of a company of the same name. Log management is a crucial aspect of software monitoring and observability, and one that often brings a surprising amount of complexity.
Though the diagram above wasn’t intended to show what log management may look like in the context of hybrid cloud, the “other environments” piece can easily be imagined as one or more on-premises data centers that a company with a hybrid cloud deployment has to maintain. Virtually any issue that arises during software development and deployment requires looking at logs generated by applications to diagnose the issue. Intuitively, having a single place to view these application logs can save time for IT teams while performing their jobs.
Another example of Anthos giving a speed boost to hybrid cloud teams is Anthos Config Management, shown in the Anthos Components diagram above in both the GCP and on-prem deployments.
From GCP’s website:
With Anthos Config Management, you can create a common configuration across all your infrastructure, including custom policies, and apply it both on-premises and across clouds.⁴
Application configs are values injected into an application either at startup or while the application is running which define bounds or variables with which the logic of the application is to operate. To give a contrived example, if you’re building a photo-sharing application, you might define a config variable that places a limit on the number of photos that a user is allowed to upload per day. You can swap in various values for this limit to change the number itself, while keeping the core application logic the same.
Anthos Config Management helps hybrid cloud teams to manage applications in public clouds and their on-premises data centers by integrating with kubernetes wherever it is deployed and giving teams a single, centralized location in which kubernetes and other configurations can be monitored and changed dynamically. Just like Stackdriver does with application logs, Anthos Config Management in the end saves time and enables teams to move more quickly.
Competitors to Google Anthos
While the content above focuses specifically on hybrid cloud deployments (splitting workloads between public and private clouds), Google Anthos can also operate on multi-cloud deployments. Multi-cloud refers to the practice of deploying applications across multiple public clouds (e.g. deploying to both GCP and AWS). Doing so helps enterprises to avoid vendor lock-in, or being too dependent on a single company as one’s cloud platform vendor.
Enabling multi-cloud means that Google Anthos makes it easier for customers to run workloads in the data centers of Google’s competitors, namely AWS and Microsoft Azure. This kind of relationship makes for interesting competitive positioning, as GCP tries to integrate nicely with other public clouds, while also competing directly against them for market share.
Of course, AWS and Azure have their own solutions to solve the same problems that Google Anthos solves. AWS’ solution is known as AWS Outposts, and Microsoft’s is Azure Arc.
Azure Arc and Google Anthos take pretty similar approaches. Both solutions attempt to be flexible in allowing customers to run their workloads in any cloud, public or private, as long as their applications are containerized and the data center is integrated with the core GCP or Azure software.
Aws Outposts takes a different approach by providing AWS hardware and software resources to a customer’s location in order to give an AWS experience and AWS APIs on-premises. For many years, AWS had touted a public cloud-only approach to enterprise computing, so its announcement of the launch of Outposts in 2018 was a surprise.⁵
Given that Anthos and Arc give flexibility in choice of hardware and type of data center, I imagine that these solutions may better serve larger customers that have already made heavy investment in their own hardware, while Outposts may better serve customers with smaller workloads that want to avoid procuring and managing new hardware.
Whichever approach is taken, the bottom-line point is the same. Large enterprises need hybrid cloud deployments, and those deployments introduce complexity in IT organizations. Solutions that reduce friction in managing such deployments will have increasing market opportunity in the next few years.
- https://www.gartner.com/smarterwithgartner/gartner-predicts-the-future-of-cloud-and-edge-infrastructure/
- https://azure.microsoft.com/en-us/overview/what-is-hybrid-cloud-computing/
- https://blogs.gartner.com/marco-meinardi/2018/11/30/public-cloud-cheaper-than-running-your-data-center/
- https://cloud.google.com/anthos/config-management
- https://iamondemand.com/blog/google-anthos-azure-arc-aws-outposts-the-race-to-dominate-hybrid-and-multicloud/
