Through advancements to back-end systems, providers of cloud-based disaster recovery as a service (DRaaS) are becoming more efficient at storing customer data and getting faster at restoring sites.
The latest company to roll out upgrades is Axcient, a relatively new player in the DRaaS market. The company says by switching to a software-based approach to managing commodity hardware, instead of expensive proprietary hardware options, it has reduced the amount of power used to back up entire business operations by 66% and has cut the capacity required to store the data by half.
Advances like these are just some of the characteristics customers should look for when evaluating a cloud-based DR provider, analysts say.
[MORE-CLOUD DR:Vendors are ready, but are enterprises?]
Founded in 2006, Axcient is a Mountain View, Calif.-based company with backing from Allegis Capital, Peninsula Ventures, Thomvest, and Scale Venture Partners, and now boasts about 3,500 customers. Many customers work with the company through a channel of more than 1,500 partners who resell Axcient on-premises appliances that are backed up to the Axcient cloud.
Axcient uses an approach that combines an on-premises appliance, which acts as a gateway between customer sites and its cloud, and manages information backed up into Axcient's back-end system. Within Axcient's cloud, the company has replaced flash storage and spinning disks with SAS drives, while networking components have been replaced with RAID cards. New software algorithms reduce the metadata needed to recover sites, which are tested with random IO benchmarks. The company, which has added four petabytes of storage to its cloud in the last year, say the single-copy of business operations are available from multiple recovery points through the cloud-based system.The company is attempting to make a name for itself in a market dominated by companies like IBM, HP and SunGard Availability Services. Gartner predicts the DRaaS market could be worth $5.7 billion by 2018.
As more providers embrace software-based approaches to managing their systems using commodity hardware, customers benefit. Outsourcing to a provider reduces capital investment needed by the customer and it provides an easier architecture for companies that have multiple sites that they want to back up to a single location.
There are still considerations that have to be taken into account though. A primary concern would be having network capacity limitations. "A key performance metric will be the rate at which you can transfer virtual machine images from your premises to the cloud," says Gartner DR expert John Morency. "A lot of that is gated by the network capacity and wide area latency." Many vendors have built-in WAN optimization solutions to reduce latency. But, companies should consider what options they have if an immediate spin up of a business continuity plan is needed.
Another consideration is that working with a vendor, especially one like Axcient, brings up the relationship between the provider and reseller. Customers may access Axcient services through a third-party reseller or the vendor. In some circumstances the reseller may be responsible for the service-level agreements and continuity guarantees, not the underlying provider like Axcient.
There are some ways to get around these issues. For the network latency challenges, users can migrate more of their workloads to the cloud, in which case the failover to the disaster system is an intra-cloud switch as opposed to being from an on-premises location to the cloud. Other vendors work more directly with customers as opposed to through channel partners.
This story, "How cloud-based disaster recovery is getting more efficient" was originally published by Network World.