Multicloud Storage

Hybrid Deployment on Google Cloud: Meet Google Anthos

People and organizations across the world have embraced the cloud. One way or another, they are leveraging cloud services ranging from infrastructure-as-a-service to fully managed software-as-a-service. However, a full-on public cloud-only technology strategy can be incredibly challenging within large corporations, due to the amount of technology landscape and culture involved. This is especially true when using hybrid or multicloud storage.

Google announced Anthos as part of their Google Cloud Platform to address IT modernization needs, including hybrid or multicloud platform deployment strategies.

There can be many roadblocks enterprises run into when trying to modernize their IT deployments. A few examples include large legacy infrastructures, regulations that require keeping data on-prem, hybrid architectures. Google’s solution for these challenges is Anthos.

In this post we’ll take a closer look at Anthos, how it can be used for hybrid and multi-cloud deployments, and the added value of Cloud Volumes ONTAP working together with Anthos.

Introducing Google Anthos

Google announced Anthos as part of their Google Cloud Platform to address IT modernization needs, including hybrid or multicloud platform deployment strategies.

Anthos is an application management platform designed for different types of workloads—brand new applications or modernizing existing ones—and it enables those workloads to run anywhere with the same consistency. Anthos was built on top of popular open-source projects such as Kubernetes, Istio, and Knative, which enables application developers to take advantage of the tooling already available for these tech ecosystems.

Enabling workloads to run seamlessly across on-premises systems and public cloud is not very common, yet is not entirely new either—several companies have interesting services in this space, such as, AWS Storage Gateway, Azure Arc, Azure StorSimple, AWS Outposts, and VMWare Cloud Foundation.

While the concept is rather similar, the technical approach varies greatly. Google Anthos was built on top of Kubernetes, leveraging its native capabilities plus extending them with some familiar Google Cloud Platform managed services such as Stackdriver (centralized monitoring, logging and metrics), Cloud Build (continuous integration/continuous delivery), Apigee (API management), and Cloud Run (a serverless approach using containers).

Kubernetes at Its Core

At the very core, Anthos was fully built based on Kubernetes. If you are familiar with Kubernetes (or simply referred to as K8s), this shouldn’t come as a surprise. After all, Kubernetes was created and developed initially by Google itself (codename Project Seven of Nine and influenced by their Borg system) and later on made available as an open source project via the Cloud Native Computing Foundation.

In a nutshell, Kubernetes enables the orchestration of containers, plus their deployment, scalability and management using automation. A container is a way to package an application independently from the infrastructure and operating system. It follows the mantra of “write once, run anywhere”, thereby providing application isolation and making it extremely portable and vendor neutral.

Container technology itself has also evolved greatly in the past few years, mostly propelled by the gigantic success and popularity of the Docker container tools.

How Does Anthos Work in Hybrid or Multi-Cloud Cloud Architecture?

Anthos leverages Kubernetes to provide a seamless hybrid experience, using the Google Kubernetes Engine (GKE) within Google Cloud Platform and the GKE On-Prem for data centers. Since it’s Kubernetes, it’s vendor neutral, so the Anthos approach can also be used in AWS and Microsoft Azure using the Google Kubernetes Engine, to maximize Anthos’s potential and multicloud platform properties. On top of Kubernetes and it’s native features, the Anthos platform provides then different layers:

  • Platform operations layer: Enables storage, container runtime and network connectivity.
  • Network and security operations layer: Enables policies, controls, and compliance across the environments (e.g. identity management)
  • Application operations layer: Enables fast and modern application development.

For a seamless developer experience in the application operations layer, Anthos leverages Google Cloud managed services and Istio, another open source project pioneered by Google. Istio is a service mesh designed to work with distributed applications and simplify development operations. It is a critical component for Anthos, because it enables visibility, security, policies and easy software delivery and roll out.

Anthos multi-layered approachAnthos multi-layered approach

This layered combination makes Anthos extremely useful for engineers who are required to use both on-premises and public cloud resources, to use the modern toolset from Google Cloud Platform, and to simplify the application development experience and seamless hybrid deployment.

Can Traditional VMs Become Containers?

Although some on-premises applications are based on container technology, the typical on-premises infrastructure isn't. When thinking on-premises, we are usually expecting virtual machines. Therefore, the key question is this: can virtual machines become containers?

Typically this wouldn’t be possible in a simple way, yet, but Google is challenging this with the release of Migrate for Anthos tool. This makes it possible to convert workloads (including physical servers and workloads that were not designed for containerized architecture) directly to containers and Kubernetes. This is a key factor to enable the adoption of this technology. However, just because in theory you can migrate any workload, doesn’t mean you should. Containers have several limitations due to the intrinsic nature of the technology and they shouldn’t be treated as a silver bullet.

What Use Cases are Good or Bad for Anthos?

The use cases where Anthos shines are the most obvious ones: workloads that you would normally run in containers and Google Kubernetes Engine but that for a given reason would benefit more from being on-premises. Of course, much more than simply hosting workloads, Anthos provides a holistic experience in application development, and that includes running workloads (on-prem and cloud) using Kubernetes.

The reasons to keep workloads on-premises may vary, but mostly such requirements are due to non-technical constraints. It could be due to security and compliance (e.g., your system needs to run in a certain location by law) or perhaps data locality (e.g., your service needs to be in the same region as certain datastore due to latency constraints).

Since Anthos is container based, it’s essentially tied up with computational workloads that can run in containers. A computational workload that requires certain specific hardware capabilities and that runs on bare metal is therefore excluded.

How about data storage systems? Well, Anthos does not really address this issue. While containers can of course have data persistency using volumes, these are not really adequate for large database systems or network file systems. To mitigate this, Anthos is available with plugin support for certain on-premises storage solutions such as the NetApp HCI, but that is of course only part of the challenge.

When addressing stateful container workloads—which require data to be permanently stored and available in case the container fails—it is crucial to understand the underlying data storage systems and how to use persistent volumes.

The data storage system is responsible for providing adequate data protection, such as backup/restore and high availability, and other data management features.

This is a scenario where a Google Cloud deployment with Cloud Volumes ONTAP and the NetApp Trident provisioner can be worthwhile. Cloud Volumes ONTAP built-in storage capabilities add advanced storage management functionality that is important when dealing with stateful container workloads. Features such as space efficient snapshots, storage cloning, high availability, disaster recovery capabilities, and data replication between clouds and hybrid deployments can make a huge difference, especially in enterprise environments and critical workloads.

Conclusion: More Than Hybrid Cloud Deployment

As more companies embrace the cloud, it’s becoming obvious that there is no one approach to the technology. This is making multicloud and hybrid platforms more relevant, and where services such as Anthos can be real differentiators.

As mentioned, Google Anthos tackled this problem by leveraging their internal known-how and expertise in Kubernetes and building bridges to other ecosystems. While not a silver bullet for hybrid deployments, it does address several key parts, namely associated with computational workloads and the seamless developer experience. For storage and overall data persistence at scale, you should explore Cloud Volumes ONTAP on Google Cloud. Also, the very same technology is available for AWS and Azure.

Cloud Volumes ONTAP was built for data management and providing top-notch enterprise-level functionalities such as storage efficiency, data protection and cloning, and hybrid deployment.

New call-to-action

Cloud Solution Architect

Principal Architect & Technology Advisor