Featured article by Roger Lam, Director of Engineering at MBX Systems
Despite the benefits of cloud backup, relatively few organizations are using the cloud as their primary data storage tier. One of the consequences is that on-premise backup appliances remain a popular solution for many organizations, particularly small and mid-sized enterprises that lack the resources to implement big-company business continuity strategies.
Steady growth in the sector has been fueled over the years by ongoing technology advances ranging from improved recovery point and recovery time objectives (RPOs and RTOs) to newer features like hypervisor support enabling instant data recovery via spin-up of virtual machines and cloud tiering allowing seamless integration between on-premise and cloud storage resources.
Even in Q3 2016, when falling storage prices contributed to declining revenues in the purpose-built backup appliance market for the first time in many quarters, analyst firm IDC reported a 12% year-over-year increase in capacity sold with a record 920 petabytes of backup storage space shipped by leading vendors.
A newer breed of backup appliance is also making its mark as part of the burgeoning Disaster Recovery as a Service (DRaaS) movement. DRaaS backup services offer the option to store data in a dedicated appliance installed at the customer site before replicating it to a public or private cloud, delivering both throughput advantages and failover alternatives. MarketsandMarkets has forecast a 45.9% CAGR in the DRaaS market from 2016 to 2021, courtesy of benefits like pay-as-you-go pricing and near-real-time failover of systems and data.
When the first integrated hardware/software backup appliances hit the market roughly a decade ago, the ability to merge backup software and onboard disk storage in a single ready-to-deploy physical unit was a game changer. It reduced the overhead associated with tape backup, provided easy setup and maintenance that eliminated the need for dedicated on-premise backup staff, and made it both faster and easier to retrieve data when needed.
Over the years, a steady string of software improvements has boosted the value of backup appliances. Data deduplication has dramatically reduced storage needs. Continuous data protection has replaced once-a-day backups with smaller and more frequent block-level backups in many systems. The need for scheduling and troubleshooting backup windows has therefore largely disappeared, along with excessive data loss caused by backups that are as much as 24 hours old.
Server virtualization has added other key capabilities. Point-in-time data or systems can now be restored by duplicating the virtual server image and running the clone on any physical or virtual server, eliminating the need for dedicated backup hardware. Data and system recovery can now be shortened with instant recovery technology, which uses hypervisor snapshots in the appliance to protect backup copies of VMs against unwanted modifications until a full restoration operation is complete. Backup and disaster recovery are converging in this and other ways.
Similarly, cloud tiering is facilitating the convergence of on- and off-premise backup by transparently moving inactive data or select workloads to the cloud for backup and redundancy. This feature enables IT teams to move inactive data to lower-cost storage tiers and free on-site storage for newer content without manual intervention. It also supports the 3-2-1 storage backup strategy stipulating that organizations should maintain two local copies of data on different devices and at least one copy off-site.
Other improvements in the past few years stem from newer hardware componentry that has paved the way for upgrades in performance, storage capacity and other functions.
The emergence of flash storage, for example, has made it possible to combine solid-state drives (SSDs) with hard disk drives in the appliance. The resulting tiered flash storage delivers faster backup processing and faster responses to data recovery requests by allowing the SSDs to function as a second-layer read cache. The most active metadata is stored on and read from the SSDs to avoid overloading the HDDs with transactional tasks that will slow performance, then automatically removed from the SSD cache when it is no longer needed.
Another major hardware advance involves the development of scale-out storage enabling the total amount of disk space to be expanded by adding new devices in connected arrays with their own resources. This modular architecture allows IT teams to buy the backup capacity they need and then add new nodes as data volumes grow, eliminating the need to purchase larger-than-needed storage arrays to accommodate anticipated future expansion that may never occur. Performance increases as storage capacity expands and multiple clusters are administered as a single system, providing important benefits over traditional scale-up systems.
Many of these advances laid the groundwork for the on-premises appliances now being used to provide local backup and temporary storage for DRaaS services.
By housing data on-site in the appliance before moving it to the cloud in blocks throughout the day, these services avoid direct-to-cloud data transfer that risks slowing or crippling the customer’s Internet connection. They also enable VMs to be failed over either to the appliance or the cloud in the event of an outage, data corruption or other service interruption.
One DRaaS provider that offers a physical appliance deployment model is Infrascale, named a Visionary in Gartner’s 2016 Magic Quadrant for Disaster Recovery as a Service in part for its 15-minute failover guarantee. If Microsoft Exchange crashes or a SQL database gets corrupted from ransomware, Infrascale can bring it back online in 15 minutes by leveraging the most recent backup on the local appliance. Customer configurations can also include both physical and virtual appliances, offering a hybrid strategy that maximizes flexibility.
As backup and disaster recovery continue to evolve, so will the appliances that support companies’ business continuity initiatives. Fewer hardware systems may be standalone in the future. More will work in tandem with the cloud. Some will morph into cloud gateways. But in one form or another, backup appliances are likely to remain a core component of the business continuity infrastructure for years to come.
Roger Lam is Director of Engineering at MBX Systems, a provider of engineering, hardware and fulfillment programs and services for independent software vendors and service providers that deploy their products on customized hardware.
This article was originally featured in IT Briefcase.