Google Cloud’s Anthos promises a single, consistent way to manage Kubernetes workloads across on-premises and public cloud environments.
Google Cloud launched the Anthos platform in April 2019, promising customers a way to run Kubernetes workloads on-premises, in the Google Cloud, and most importantly, in other major public clouds, including Amazon Web Services (AWS) and Microsoft Azure. Speaking at Google Cloud Next in San Francisco in 2019, Sundar Pichai, CEO of Google, said the idea behind Anthos is to let developers “write once and run anywhere”. – a promise to simplify the development, deployment and operation of containerized applications on hybrid clouds by bridging incompatible cloud architectures.
It took some time to get this crucial multi-cloud support. Google finally announced support for Anthos for AWS (from April 2020) and Anthos for Azure (from December 2021), with the release of the Anthos Multi-Cloud API, fulfilling its initial promise of a true hybrid and multi-cloud operability.
The Google Cloud Anthos console showing Azure and AWS assets. (Credit: Google)
By providing a single platform for managing all Kubernetes workloads, Anthos helps customers focus their skills on a single technology, rather than relying on certified experts in a host of proprietary cloud solutions. Similarly, Anthos provides operational consistency between hybrid and public clouds, with the ability to apply common configurations across all infrastructures, as well as custom security policies tied to certain workloads and namespaces, regardless of where these workloads are executed. Finally, IT operators can track cluster telemetry and log information from a single console.
A multitude of components
Anthos is the natural evolution of the cloud services platform the vendor was building before 2019. The platform combines the Google Cloud managed service, Google Kubernetes Engine (GKE), GKE On-Prem, and the Anthos Config Management console for administration, unified policies and security across hybrid and multi-cloud Kubernetes deployments. If we add Stackdriver for observability, GCP Cloud Interconnect for high-speed connectivity, Anthos Service Mesh (based on the Istio open source project) and the Cloud Run serverless deployment service (based on the Knative open source project) , GCP seeks to provide a one-stop-shop for managing Kubernetes workloads, regardless of where they are produced.
Based on GKE, Anthos automatically takes care of Kubernetes updates and security patches as soon as they are released. Native installations of GKE On-Prem will run on VMware vSphere or bare-metal, with launch partners VMware, Dell EMC, HPE, Intel and Lenovo committing to deliver Anthos on hyperconverged infrastructures.
Compete with AWS, Oracle or even Microsoft
The fear of being blocked by a supplier is very real for companies. Providing a flexible and open path to migrate to the cloud is something of a holy grail for cloud providers today. But some want to have butter and butter money, locking those customers into their own ecosystem when they decide to move workloads to the cloud. Amazon Web Services finally gave in on the hybrid cloud front by announcing Outposts to help customers bridge the gap between on-premises and cloud workloads. Thus, Outposts combines hardware configured by AWS and services and APIs managed by it. Then, in December 2020, AWS extended its Amazon Elastic Kubernetes (EKS) managed service to workloads running both on-premises and in the AWS Cloud.
Oracle Cloud at Customer and Microsoft Azure Stack are similar hybrid cloud offerings offered by other major players, while platform-as-a-service offerings Red Hat OpenShift and VMware Tanzu, both powered by Kubernetes, enable containerized enterprise applications to run in hybrid and public clouds. In an attempt to dethrone these big rivals, Google Cloud is betting big on the fact that Kubernetes is the future of enterprise infrastructure. Of course, competitors are also pushing aggressively into the world of Kubernetes management, but as the petri dish where Kubernetes was developed, Google has strong claims to be the best way to run this technology.
A simplified migration to Anthos
To help customers get started, Google launched Migrate for Anthos following the 2018 acquisition of Velostrata, an Israeli company that specializes in migrating to the cloud by intelligently decoupling storage and compute. So businesses can leave storage on-premises and run compute in the cloud. Migrate for Anthos converts workloads into containers for Kubernetes directly from physical servers and virtual machines. How does it work? Migrate for Anthos scans the file system of a server or virtual machine and converts it to a Kubernetes persistent volume. Application containers, service containers, networking, and persistent volumes come together in a Kubernetes pod, which is a group of containers deployed together on the same host.
For GCP customers, getting started with Anthos is as simple as creating a new GKE cluster, with the Istio service mesh enabled, in the console. For on-prem customers, the first step in using Anthos is to set up a GKE On-Prem cluster and migrate an existing application. Once this cluster is registered with GCP, it is enough to install Istio to obtain visibility of the workload on all the clusters. Then, by enabling Anthos Config Management on GKE clusters, all Kubernetes and Istio policies can be managed in one place.
Google Cloud Anthos pricing
Google Cloud Anthos is available on a pay-as-you-go or monthly subscription basis, with discounts based on commitment. For cloud customers, Anthos costs $8 per cluster vCPU per month on a pay-as-you-go basis, or $6 if paying on a subscription basis, regardless of the public cloud platform on which the workload works.
For on-premises customers, the cost of Anthos is $24 per cluster vCPU for pay-as-you-go customers running on VMware or bare-metal. A free trial for new customers allows the use of up to $800 for up to 30 days. (Prices are not available in euros).