What is Scale-Out Architecture?

Posted 05/16/2018 by Greg Page

You need Scale-out in your backup

As George Crump accurately noted in his blog, "Why Backups Break and How to Fix It", there are many issues in the current backup landscape, including legacy backup applications, fragmentation of tools to support new technology and decentralized management of backup for applications, the cloud, storage and virtual platforms.

The question, then, is how do you solve the problem of protecting and managing your data in a world where those applications and data can dynamically grow, change size, location, platforms or virtualization layer? The solution is easier than it may appear, as the problem resides in the underlying technology: scale-up.

Challenges with Scale-up Technology

In a traditional scale-up model, you typically see servers, network and storage provisioned, managed, upgraded and retired as separate items. Also, they are often purchased or leased on different schedules and with different capabilities. That leads to forklift upgrades and time-consuming expensive migrations to support that upgrade effort. No one enjoys trying to plan the timing of upgrading or replacing servers to match storage on different refresh cycles while moving data, applications and users around in the backup environment to avoid gaps or outages in services.

To add to this complexity, consider that you may be doing this for multiple backup products or appliances with varying feature sets and application support. Then to top it off you are supporting aggressive RPO/RTO service levels, traditional and mobile users, with applications that need to move seamlessly between the cloud and on-premises, which then makes backup and data management look very bleak.

What do you do to solve the challenges? Leverage the benefits of scale-out technology.

The benefits of Scale-out Technology

You cannot continue to scale-up for data protection. Instead, you need to examine how scale-out technology addresses these challenges.

A scale-out cluster can be loosely defined as a series of co-equal nodes working together. As nodes are added and removed, throughput, compute and bandwidth increases or decreases accordingly. Upgrades are accomplished by adding new nodes to the cluster, seamlessly moving users, workloads and data to the new nodes within and retiring the old nodes with no downtime. A characteristic feature is the inclusion of machine learning to facilitate self-healing and auto-tuning of resources in the cluster to accommodate changing demands. Scale-out systems, with the proper design, are also well suited to accommodate cloud use cases and run in the cloud.

It’s not hard to see the advantages of a backup system with these capabilities. Converging your backup environment into a single compute, network and storage cluster with native cloud and hypervisor integration has a distinct value for any organization.

Combining this approach with broad application, cloud, and hypervisor support, multi-tenancy, integration with more than 20 storage vendors with value-added features like archiving, sensitive data governance - and REST capable interfaces and management - solves the issue of broken backup.

Commvault HyperScale™ Technology brings together the Commvault platform with scale-out, hyperconverged technology to create a flexible, predictable and less expensive backup and data management solution.

There is a solution for all of the things wrong with backup. They don’t have to be broken any longer.