Backup is an essential part of a disaster recovery and business continuity strategy, yet building backup infrastructure for more than even the smallest use case is no simple task.
The traditional data protection approach is to source and deploy physical servers and then move the backup software onto the hardware. But this process requires having team members who have the design skills necessary to understand the parts of the architecture that impact throughput and how the components of the backup software work together.
An increasingly popular approach is to use a dedicated data backup appliance. An appliance removes much of the heavy lifting involved in designing and implementing a backup architecture. While not suitable for all deployments, many IT organizations can benefit operationally -- and perhaps financially -- from implementing an integrated backup appliance.
Appliances for data backup actually have existed for many years. However, in effect, these were simply repositories for data that optimized the content stored on them. Virtual tape libraries and other platforms that offer interfaces, such as network file systems, typically do not include backup software. Instead, they simply deduplicate, compress and store data for future use. Although this serves a purpose, separate backup software is still necessary.
An integrated backup appliance eliminates the need for backup software. The appliance works directly with the application, server or hypervisor to protect virtual machines (VMs), databases and other application data. The appliance delivers all functions needed for future data restore, such as cataloging, indexing and search.
The newer generation of integrated appliances is billed as converged secondary storage, handling all data management beyond primary storage. These appliances include functions such as backup, archiving, disaster recovery, copy data management and test/dev data.
But the appliance model is not for everyone. For example, it may be more practical for SMBs to use cloud-based backup or a product that writes to a local hard drive. However, once environments scale, especially with server virtualization, appliances start to hit their sweet spot. The custom model is potentially more practical for very large enterprise environments, as requirements at that level tend to be more bespoke.
In that middle ground, integrated appliances could be useful for the following circumstances:
- Organizations with scalability or management issues. If backup is taking an inordinate amount of management time, an appliance model may be the right choice. Scaling traditional backup can be a problem, so an appliance may help with the operational and financial aspects of scalability.
- Organizations with no dedicated backup team. Because not everyone can afford a dedicated backup team, backup is a shared responsibility in most IT organizations. An integrated backup appliance can minimize or eliminate some operational aspects of backup, greatly reducing operational workload.
Most large enterprise organizations run multiple platforms across data centers, including mainframe, Windows and Unix-based systems. For these groups, self-build and design is probably more appropriate, as it offers the most flexibility and degree of additional control over the hardware and software.
Using extensive research into the data backup market, TechTarget editors focused on integrated backup appliances that include hardware and software that are provided by the same vendor. Our research included data from TechTarget surveys, as well as reports from other respected research firms, including Gartner.
Identifying data backup appliance use cases
Appliances are great at serving what could be classified as general backup use cases. Typically, this means the ability to create short-term and long-term backups depending on service-level requirements. Most integrated backup appliances implement some form of data deduplication, as this significantly reduces the volume of data stored. Because most IT shops are highly virtualized, deduplication provides the ability to significantly reduce the volume of physically stored data.
This makes the appliance model highly suitable for the retention of long-term backups that can also serve as a form of archive. The data backup appliance needs to simply keep track of differences between each backup, effectively implementing "incremental forever" backups. That means only the incremental differences are retained after the first full backup copy of a VM or system. From an application perspective, backups taken of VMs are not really archives in the strictest sense. They do, however, provide a timeline of application changes that are relatively easy to recover from and are good enough as an archive platform.
Should I choose an appliance?
If your organization is considering a backup appliance, ask yourself the following questions:
- Am I exceeding my backup windows?
- Am I investing too much effort in building and managing backup products?
- Do I want more automated and self-service solutions?
- Could I justify secondary data usages for backup content?
If the answers to these questions are more yes than no, then a data backup appliance may be a good fit.
Many appliance vendors have implemented continuous data protection functionality, ransomware protection and self-service through software in their platforms. As most data backup appliances have an underlying file system into which backups are stored, many of these features are easier to implement than on traditional backup systems. That takes us into a bigger discussion of how appliances are differentiated.
Some of the features discussed here could equally apply to traditional backup environments, so how are appliances different?
- Turnkey. Appliances are deployed as a "black box" that does not have to be assembled from component parts. The vendor does the integration, selecting the right server and storage components that will work with the backup software.
- Ease of deployment. The customer simply buys the data backup appliance, racks it and powers it on. The aim here is to remove the complication of deployment, which should take minutes or hours instead of days.
- Vendor-managed. As part of a turnkey system, vendors manage the product from an infrastructure perspective. That means they replace failing components and manage software upgrades and patches. As a single source of all data in the enterprise, the backup system is an obvious asset for hackers to target, so being confident the system is fully patched is important.
- Single licensing model. Typical licensing models are based on either capacity or the cost of individual nodes. Features are usually bundled in the cost and are not itemized separately.
- Single point of truth. Using a backup appliance -- especially a scale-out product -- enables organizations to create a single source of truth on backup data. This is true even when the physical appliances are dispersed in branch configurations because the metadata for backups is still stored centrally.
- Cloud extensibility. Many products now support the extension of backups to the public cloud. Initially, these offerings just used the cloud as a backup repository, but now, there are products that can run virtual appliances in the cloud and gain access to the backup data for restoring VMs or performing workload migrations.
- Secondary data usage. A key benefit of the backup appliance model is that it can be used as a host for active data. With server virtualization, the appliance can act as a data store and be used to start VMs based on backups. This is a powerful way of recovering data or proving the validity of backups as they are taken.
Disadvantages of the data backup appliance model
Naturally, the appliance model is not right for every situation, and there are times when an integrated backup appliance may not be as practical. The first is in support. Today, appliances have been generally developed by startups, so there may be limited support for all platform types, including older Linux systems, mainframes and bare metal. The relevance of the range of platforms supported comes down to how diverse an IT environment is. Larger enterprises may find that their use of many legacy platforms makes appliances harder to adopt, unless they are more mature in platform coverage. When we examine the vendors, we will identify products on both sides of this problem.
The next issue is lock-in. Highly deduplicated data is great for saving on physical storage. However, the dedupe process is proprietary, and backup images must be rehydrated or copied out of the backup system before they can be moved to a new platform.
Backup platform migration is a tricky problem to solve, but it can be at least mitigated by having as little long-term retained data as possible. Backups can then "roll over" pretty quickly as the backup process is moved to another system. The implications of backup policy design are outside the scope of this article but are relevant when deciding which data might be appropriate for an appliance-based product.
Finally, there's a question of the vendor's hardware margin. Larger organizations may already have terms in place with vendors for low-cost servers, resulting in paying a margin for acquiring a hardware-based backup appliance. Bear in mind: This margin does cover the testing and certification done by the vendor, so it does have some value.
The market for integrated data backup appliances
As we delve further into backup appliances, we will examine the overall market and focus specifically on the companies and products viewed as market leaders. This includes a division of new and old; newer startups, like Cohesity and Rubrik, are challenging incumbents, such as Arcserve, Barracuda Networks, Commvault, Dell EMC, Unitrends and Veritas Technologies. We will look at whether the architecture of new platforms can deliver a differentiation over the wealth of experience the more established companies have to offer. Is the appliance model simply a repackaging exercise, or do these companies have something new?