Google Cloud has announced the general availability of Anthos: a platform that allows customers to run applications on-premise, in the Google Cloud and, crucially, with other major public cloud providers like Microsoft Azure and Amazon Web Services (AWS).
Speaking at Google Cloud Next in San Francisco yesterday, Google CEO Sundar Pichai announced the idea behind Anthos it is to "write once and run anywhere”.
“It gives you the flexibility to move on-prem apps to the cloud when you are ready, and it allows you to keep using the technologies you are already using, while improving security," he said.
"Today most of the world's enterprise computing still happens on-premise,” he added. “It hasn't moved to the cloud yet because the path forward is complex and daunting, full of difficult decisions. How do you modernise without jumping completely to the cloud? How do you bridge incompatible architectures while you transition? How do you maintain flexibility and avoid lock in?"
The previously released Google Kubernetes Engine (GKE) and GKE On-Prem allowed for hybrid deployments, yet customers continued to demand a platform that made it simple to span multiple, rival cloud providers as well.
Nick McQuire, vice president of enterprise research at CCS Insight said: “With the arrival of Anthos, and in particular its support of open source, particularly Kubernetes, Google is now taking a much more realistic path in meeting customers where they are on their cloud journeys and is aiming to become the standard in hybrid, multi-cloud services in this next phase of the cloud market.”
New Google Cloud CEO Thomas Kurian agreed: "Anthos came from listening to customers who wanted three important things from their cloud providers: First, a technology stack to run in their data centre next to enterprise workloads that they couldn't yet move to the cloud; Second, a single programming model that gave them the choice and flexibility to move workloads to both Google Cloud and other cloud providers without any change; Third, a platform that allows them to operate this infrastructure without complexity and to secure and manage across multiple clouds in a single, constant way."
Speaking about customer benefits at a press roundtable later that day, Kurian again demonstrated a politician's propensity to group ideas into threes: "One, because of the scarce skills in learning cloud, it helps them train on one technology. You don't have to learn every cloud provider's technology; You can learn one technology and use it consistently everywhere. Second, it gives them choice. They don't get locked into any cloud provider; They can choose to migrate to whichever cloud they have the best solution from.
"Third it gives them operational consistency, which is a very important thing. For example, security can be configured the same way, monitoring can be configured the same way, policy can be configured the same way - which is very important for customers, because otherwise it adds risk."
Fear of vendor lock-in is very real for enterprise customers. Providing a flexible and open route to move to the cloud is something of a holy grail for cloud vendors today. But some want to have their cake and eat it, by trapping those customers within their own ecosystem when they do decide to move workloads to the cloud.
Amazon Web Services announced AWS Outposts late last year to help customers bridge the on-prem and cloud world as seamlessly as possible. Oracle's Cloud at Customer and Microsoft Azure Stack are similar hybrid cloud offerings from other major players.
For example, with Outposts customers receive AWS-configured hardware and software delivered to their on-premise data centre or co-location space. Here it can run applications in a cloud-native manner without having to touch AWS' data centres, which could provide a simpler route to 'lifting and shifting' apps from existing infrastructure into AWS' cloud - but not to rival public cloud providers such as Azure or GCP.
With Anthos, Google is setting out its stall to be the only truly flexible option on the market.
The platform is designed for fluidity, explained SVP of technical infrastructure Urs Hölzle and VP of engineering at Google Cloud Eyal Manor in a Google Cloud blog post: "Anthos will also let you manage workloads running on third-party clouds like AWS and Azure, giving you the freedom to deploy, run and manage your applications on the cloud of your choice, without requiring administrators and developers to learn different environments and APIs."
On the topic of skills, Kurian said: "Many large companies truly want to train the people once on a set of technology that they can deploy anywhere. None of the other cloud providers are solving that.
"Today, if you talk to Azure they will say you can run Azure Stack on-premise and on the cloud, Amazon will say you can run Outposts on-premise and in the AWS cloud. They are fine companies, but they're not solving the multi-cloud problem.”
Under the covers
Anthos is the natural evolution of the Google cloud services platform announced last year. It brings together a combination of the existing Google Kubernetes Engine (GKE), GKE On-Prem and the Anthos Config Management console for unified administration, policies and security across hybrid Kubernetes deployments.
Anthos "is a 100% software-based solution," Hölzle’s blog post said, boasting that it’s easy to get up and running on existing hardware with no forced stack refresh, and that because it is based on GKE, Kubernetes updates and security patches are added automatically as they are released.
Anthos is also hardware agnostic, so it can run on customers’ existing servers. It also launched with a number of hardware partners, including commitments from VMware, Dell EMC, HPE, Intel, and Lenovo to deliver Anthos on their hyperconverged infrastructures.
To help customers get started, Google launched Anthos Migrate in beta today. This service "auto-migrates VMs from on-premises, or other clouds, directly into containers in GKE with minimal effort,” according to the blog post.
For net new customers, the first step to running Anthos involves setting up a GKE On-Prem cluster and migrating over an existing application. Once this cluster is registered with GCP, you can install the Istio service mesh to achieve workload visibility across all your clusters. Then, by enabling the Anthos Config Management across your GKE clusters, all Kubernetes and Istio policies can be managed in one place.
Anthos is charged as a monthly term-based subscription with a minimum one year commitment. It’s then priced on incremental blocks of 100 vCPUs, starting at $10,000 per block regardless of where that workload is running.