BlueXP Blog

Google Hybrid Cloud with Anthos and Cloud Volumes ONTAP

In the first months of 2019, Google took the cloud world by surprise with the announcement of Google Anthos—a service that enables a holistic application development experience across different environments and sets the stage for a Google hybrid cloud storage offering.

Having a clear and focused offering for hybrid and multicloud architectures and deployments weren’t very common at that point and they weren’t highlighted by any public cloud provider. The other major hyperscalers, Microsoft and Amazon Web Services, followed Google’s lead and today are providing similar services such as Azure Arc and AWS ECS/EKS Anywhere.

Anthos and these hybrid solutions showed a major shift towards accommodating the hybrid and multicloud deployment model needs for organizations with complex solution architecture requirements. However, for enterprise-level cloud data storage management features and data persistency at scale, Google Cloud users may find an added value in deploying Anthos alongside Cloud Volumes ONTAP on Google Cloud.

In this blog we’ll see how Anthos and Cloud Volumes ONTAP for Google Cloud can combine to simplify data management across environments, with the added value of NetApp features in Google hybrid cloud environments.

Click below to jump down on the sections on:

Data Storage in Google Hybrid Cloud with Anthos

The main selling point of Anthos is the ability to provide engineers a unique and consistent development experience that enables them to modernize and run their workloads anywhere: on-premises, Google Cloud Platform, or any other public cloud. Anthos is aimed at hybrid or multicloud architectures and deployment models, i.e., environments that use a mix of on-premises and multicloud, and with orchestration between both of them.

To achieve this, Google leverages multiple existing services such as Google Kubernetes Engine (GKE), Istio, and Apigee. In fact, Kubernetes and containers are the key to the solution and Google went as far as creating a tool to help customers to convert virtual and bare metal servers into containers to run on top of GKE. The storage of the workloads is limited to the options GKE provides for both the cloud and the on-premises services.

Google Anthos has a very clear selling point: a holistic and consistent development experience across hybrid environments, plus the appealing message of helping enterprises to modernize existing workloads by migrating them to Anthos.

From a computational and network point of view, it’s quite easy to understand the simplicity of placing apps in containers and running them in any environment as easily as a common cloud deployment. However, one of the biggest drivers and business use cases for hybrid cloud is precisely about data locality and access to restricted network resources. So, how does storage and data persistency work with Anthos?

Since Google Kubernetes Engine (on-prem and cloud versions) is at the core of Anthos, it is implicit that the way storage works is inherited from the core Kubernetes and container runtime (earlier primarily Docker, nowadays typically containerd) capabilities. A multicloud Kubernetes deployment has many different use cases and best practices.

A container runtime such as Docker or containerd uses the concept of volumes to provide a mount point that containers can use for persistence data. Since the containers run in GKE, this request is fulfilled by whichever enabled storage provisioners and plugins are associated with the Kubernetes cluster configuration (by default the local disks in the cluster). This can of course vary from generic plugins such as NFS and iSCSI to enterprise-grade providers such as NetApp Cloud Volumes ONTAP/Trident.

While it is a good thing that Anthos uses GKE to manage storage, it also means that the same challenges and limitations of container runtime and Kubernetes storage are present in Anthos. Some of those pain points are of course very well-known such as:

  • Resilience: Container orchestration solutions were initially focused on operating stateless workloads that could easily tolerate container failures and restarts. While the technology has matured a lot in the last few years, there are still inherent challenges in running big stateful workloads (e.g., databases).
  • Availability: Enterprise storage availability features such as cloud high availability, cloning and replication are not natively supported.
  • Location: Moving data between different locations requires advanced custom logic.

Using Cloud Volumes ONTAP with Anthos: More Hybrid Cloud Benefits

If you’re using Google Cloud, storage limits imposed by Kubernetes are going to be a factor you need to look into. A good solution for these limitations is to take into use a cloud-based, software-defined storage product such as the NetApp Cloud Volumes ONTAP.

As a Google Cloud Technology Partner, NetApp actually made that compatibility crystal clear by announcing a strategic partnership with Google that includes the recognized Anthos storage support and compatibility. NetApp is Anthos Ready storage partner, a restricted qualification that Google issues for partners that can provide you a seamless storage experience.

When combined with Kubernetes (via the Trident provisioner for dynamic persistent volumes), Cloud Volumes ONTAP enables customers to immediately gain access to a wide array of enterprise level features.

Having Cloud Volumes ONTAP with Anthos immediately solves a lot of the big pain points related with the resilience and availability of data in Kubernetes such as high availability, cloud storage security and data protection (server-side and in-transit), and of course, leverage storage efficiency that come natural from having a cloud-based software-defined storage. Without it, engineering teams would have to spend a good amount of time coming up with special tools and scripts to deal with storage volumes operations such as replicate, adjust, sync, etc.

The data location challenges are also solved out-of-the-box due to the fact that Cloud Volumes ONTAP uses a native cloud approach. With both GKE on-prem and GKE in the cloud using Cloud Volumes ONTAP as a storage provider, data can be moved seamlessly between public cloud and on-prem NetApp storage systems using the SnapMirror® data replication feature. This allows proper failover and failback across environments or even disaster recovery as well as backup and recovery capabilities.

One example scenario could be a relational database that you need to be available both on-premises and in the public cloud for different use cases. Within your on-premises environment the database could be heavily used by your applications to store customer-related data, and in the public cloud you might have the need to access that database using a business intelligence tool for customer analytics and for research and machine learning training.


Google Anthos Platform using Cloud Volumes ONTAP as the storage backend.

Having this scenario would enable the same data to be available for use both on-premises and in Google Cloud. This gives a huge amount of flexibility in the location where you want your data to be available but also in the way you manage your data across hybrid environments.

Conclusion

When adopting a hybrid or multicloud strategy with a solution like Google Anthos it is paramount to have good storage capabilities. After all, data and stateful workloads are the hardest parts to figure out in that equation.

When dealing with stateful workloads such as databases or any other application that requires data persistency using NetApp Cloud Volumes ONTAP can make a big difference used in combination with Google Anthos, adding features such as storage efficiency features that can reduce cloud data storage costs, storage snapshot data protection, and instant data cloning capabilities, all of which has made it very popular in enterprise environments.