Evaluating Data Protection for Hyperconverged Infrastructure

If you're looking at deploying hyperconverged infrastructure, data protection shouldn't be an after-thought.

By Jim Whalen, Senior Analyst and Consultant, The Taneja Group

Hyperconvergence is a still-evolving trend, and the number of vendors in the space is making the evaluation of hyperconverged infrastructure complex. One criterion to consider in any infrastructure review is data protection.

Effective data protection —backup and restore, replication, disaster recovery— in the hyperconverged infrastructure space is just as important as it is in the more traditional discrete component IT space. Historically, there have been two ways that hyperconverged infrastructure vendors have chosen to provide data protection to their customers: either partner with third-party data protection software vendors or layer some data protection capabilities onto the platform so customers can go with a ‘one box’ solution if so desired. But hyperconverged infrastructure is unconventional, so if you’re thinking about it, don’t let data protection be an afterthought.

SimpliVity is taking a different philosophical approach than other hyperconvergence vendors. Whereas traditionally, hyperconvergence was defined as integrating compute, storage and networking, SimpliVity chose to integrate the entire IT stack, also pulling deduplication, compression, WAN optimization and data protection into its implementation of hyperconvergence. No one had done this before, and it now introduces a third way of providing data protection to customers – build it in from the start as a key element of the architecture to make it as simple, efficient and effective as possible.

The earliest market entrants started with factory pre-configured rack programs, referred to as converged infrastructure, to make it easier to purchase, install, and manage their equipment. Over time, the degree of integration increased, and the industry then began referring to VM-centric hyperconverged architectures, where compute, storage and networking are all tightly pulled together in one box below the hypervisor and provide an integrated, virtualized platform to run workloads on. The big benefit of hyperconvergence is that it basically provides “infrastructure in a box” that enables very short time-to-value. The administrator buys an appliance, plugs it into the IT center, and spins up the desired VMs. Scale-out is transparent; if more compute/storage is needed, another appliance is simply added to the mix, providing more IOPS and storage space to run additional VMs. Almost everything is self-managed by the appliances, significantly reducing the amount of IT involvement required. In essence, hyperconverged systems allow users to operate up at the VM level instead of down at the IT component level.

Partner With Third-Party Data Protection Software Vendors

Pivot3 and Gridstore are examples of vendors that have taken the first approach to data protection and rely solely on third-party software packages for all data protection capabilities. There are advantages to this approach. It allows the platform vendor to focus on what is presumably their core competence —providing a reliable, highly functional, converged platform. It gives the customer a choice of data protection software, which might be particularly valuable if they’re already using that software elsewhere in their data center. It also allows customers to leverage the entire third-party data protection industry to provide comprehensive, advanced functionality and features more quickly than possible on their own.

This is certainly a viable way of supplying data protection; there are a number of excellent, full-featured data protection packages out in the market. However, it costs substantially more for the extra software and dilutes the ease-of-use message of hyperconverged platforms by forcing you to install and manage a critical piece of technology that’s missing from an otherwise complete solution.

More subtly, it surrenders control of the experience while also complicating it. Now, you’re forced to work with both the management interface for the platform and the backup software. You have to learn how to configure and run the software. If something goes wrong, you have to determine if it’s the platform or the backup package. Which vendor do you call?

You may also have to deal with performance impacts to production workloads when the backup software is running. While this third-party approach to providing data protection is functional, it may not be ideal.

Layer Some Data Protection Capabilities onto the Platform

Nutanix and Scale Computing are examples of hyperconverged infrastructure vendors currently using a layered-on approach. Initially, Nutanix relied solely on third-party application vendors for all of its data protection capabilities, and today, they continue to partner with and recommend outside data protection vendors. Recently, however, Nutanix added its own backup, replication and scheduling functionality to its platform. They do this by utilizing the VM snapshotting interface on the hypervisor they’re running. They can replicate these snapshots elsewhere in their cluster or to another site, create clones and perform failover, providing native disaster recovery functionality. Additionally, they can utilize the public cloud as a backup destination. With these features, Nutanix now offers enough data protection capabilities on its platform that users no longer have to buy additional software.

Similar to Nutanix, Scale Computing can take VM snapshots, replicate them elsewhere in the local cluster or to another site, and spin them up as clones.

Build Data Protection In as an Element of the Architecture

It’s fair to ask “Why is designing data protection into the architecture from the start any better than adding it onto an existing platform?”

Well, think about renovating a house. If it’s a ranch-style home that was originally built on a slab, and the owner decides they want a basement and a second story, they have big problems. You can certainly add those things, but it’s not easy. And the results probably aren’t going to be as good as if it were designed and built that way in the first place. It’s likely to take longer and cost more, too.

This analogy is directly applicable to hyperconverged infrastructure platforms. You can simply do things better when you look at them holistically instead of after the fact. For example, as data is ingested by SimpliVity hyperconverged infrastructure, it is automatically globally deduplicated and compressed, and remains that way forever – no rehydrating, no post processing, no specialized WAN optimization when shipping anything across the wire. This means that backups are always performed on highly optimized data, so they’re only done as full backups, with no need for incrementals—ever. This makes them quick, light touch and highly space-efficient.

When SimpliVity created its platform, one of the design goals was to “abstract the policy from the infrastructure,” meaning that they wanted the user to always be operating at a high level, not worrying about the nuts and bolts of how things work. This principle is best illustrated by looking at how you add a hyperconverged infrastructure node to a SimpliVity Federation. You merely have to plug it into the network, and it’s automatically recognized by the rest of the members and starts running. There’s no user-level configuration that needs to be done —no IP addresses to set, no storage pools to map, no other sites to inform about the addition, etc. It just works. With backup, this means that it becomes an easy and high-level process that seamlessly fits in to the existing workflow. To protect a VM, for example, you merely specify how often you want to make a copy of it, how long you want to keep the copies and where you want to save them. If you migrate a protected VM to another appliance, the backup policy migrates with it. There’s no setup, configuration or scripting required.

So, this new approach to hyperconverged data protection offers a lot of upside, but what’s the downside? There isn’t any downside that’s specific to data protection being integrated in to the product architecture instead of being an add-on; integration is all positive.

Instead, the downside to having integrated data protection is the same as having add-on data protection: If there’s some feature in some backup package that you really want, but it doesn’t show up in the platform product roadmap, you may be out of luck. Another is, as mentioned above, when you’re depending on one vendor to provide everything, you might not get it all right away. For example, when SimpliVity released its initial product, it did not have the ability to do file- or folder-level restore; it was only a full backup and full recovery only. In the current release, file-level recovery functionality is now in place.

Effectively, by integrating data protection in to its hyperconverged platform, SimpliVity is pulling secondary data workloads into its architecture, blurring the long-standing difference between primary and secondary storage and carving out an even larger market for itself. After data protection, there’s still DevOps and data analytics that could be integrated. SimpliVity, with its always-on global deduplication and highly efficient copies, is well positioned to do it all. It will be interesting to see how SimpliVity’s architecture evolves to encompass these use cases.

Choosing a Data Protection Strategy

With this new paradigm introduced by SimpliVity, there are now several very different strategies in play in the hyperconverged infrastructure space. Certainly, each approach will have its adherents, but if you’ve decided that you don’t want to incur the extra expense of third-party backup software and the capabilities that come with the platform meet your needs, you need to look at SimpliVity and its built-in data protection.

So, which approach is best for you?

Short answer: it depends on how much of a pain point data protection is for your organization, and your intention to use hyperconvergence to help address it.

However, just as the deployment of virtualization in your data center in the last few years was disruptive for data protection, hyperconverged infrastructure is another catalyst event that could be a forcing function for you to reconsider your data protection approach. The simplification of infrastructure, associated services and operations may allow you to also simplify your data protection. If you’re evaluating hyperconverged infrastructure solutions, just don’t let data protection be an afterthought.

NOTICE: The information and product recommendations made by Taneja Group are based upon public information and sources and may also include personal opinions both of Taneja Group and others, all of which we believe to be accurate and reliable. However, as market conditions change and not within our control, the information and recommendations are made without warranty of any kind. All product names used and mentioned herein are the trademarks of their respective owners. Taneja Group, Inc. assumes no responsibility or liability for any damages whatsoever (including incidental, consequential or otherwise), caused by your use of, or reliance upon, the information and recommendations presented herein, nor for any inadvertent errors that may appear in this document.

Photo courtesy of Shutterstock.