BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Disaster recovery is becoming increasingly important to organizations of all sizes. But with DR budgets increasing...
only slightly, IT planners are looking at the cloud as a way to lower their disaster recovery costs. While the cloud offers some excellent capabilities for disaster recovery, is it truly a cheaper alternative than an in-house approach?
There are several cloud disaster recovery options available. The most common approach is using the cloud as a backup target. In the event of a disaster, the organization relocates to another facility and either begins downloading (restoring) the backup data or waits to have the data shipped to them on hard drives or tape.
The value of the cloud in this situation is that it provides an off-site copy of data that's accessible from virtually any location. More importantly, that location does not need to be defined in advance. Businesses avoid investing in a second site or paying for that site until there is an actual disaster.
The downside to this approach is the time it takes to recover data from the cloud. While data deduplication improves backup speed, it has virtually no value during the recovery process. For this reason, most IT planners are better off having data physically sent to them via overnight or same-day delivery.
The challenges associated with recovering data from the cloud created a second market: cloud disaster recovery-as-a-service (DRaaS). Cloud DRaaS allows an organization to run instances of its applications in the provider's cloud. The obvious advantage is that the time to return the application to production, assuming networking issues can be worked out, is greatly reduced because there is no need to restore data across the Internet.
However, there are some disadvantages to this approach, most notably the performance of applications running in the cloud. There are also potential migration issues when returning the application to the subscriber's internal data center.
Secondary site costs can add up
The obvious factor in reducing a cloud provider's disaster recovery costs is the elimination of the secondary data center. Many organizations don't have a secondary site. For those that do, whether it is suitable as a DR site is another question. Making the second site DR-ready may require new IT staffing and upgrades to the network, storage and server infrastructures.
Time-to-recover expenses can gobble cash
Another DR cost consideration is the time to return to operations. When organizations use the cloud just for backup storage, the time to recover to the DR site can be hours or, in some cases, days. This does not include the time required to find a DR site and get servers, networking and storage shipped in.
The second option, DRaaS, provides faster recoveries and eliminates the need for emergency hardware shipments. However, some providers won't commit to anything less than 10-hour recovery times and only allow companies to run a few applications during the disaster.
Some providers also charge the subscribing organization an additional hourly cost while the application is running in their cloud. A prolonged outage of 30 or more days could get quite expensive.
Long-term storage can get pricey
The big cost of cloud disaster recovery is the price of storage. While the cloud is almost always cheaper initially, it's the recurring bills that may make it more expensive over time. The longer data is stored in the cloud, the more expensive it becomes. If an organization is going to use the cloud as an archive for years of backup data, it can become more expensive than on-premises storage.
The best value would be to use the cloud to store only the latest copy of data, the version most typically needed in a disaster. The working data set grows much more slowly than the total data set, which keeps cloud storage costs under control. Making sure that only active data (data that has been accessed in the last 30 days, for example) is in that working set could reduce this cost even further.
Is cloud truly the lower priced option?
In the battle over disaster recovery costs, the cloud clearly has an advantage if there is no secondary site that's DR-capable. Typically, the costs to equip and maintain a second site overshadow even the ongoing costs of a cloud alternative. This is especially true if the organization can put up with an outage long enough to use a cloud backup option.
DRaaS is a viable option for those organizations that need a quicker recovery. But the organization needs to be aware of the following important considerations:
- How long it will take to recover an application
- What the performance of that application will be like during the disaster state
- The costs associated with running its application in the provider's cloud
However, the real stumbling block is storage. The cost to pay for the same 10 TB of data storage, every billing period for years, can be expensive. Move that capacity number to 100 TB or 1 petabyte, and the ongoing cost of cloud storage could become so challenging that it may be less expensive to keep DR in house.
The rise of cloud-based disaster recovery
Ways to make cost-effective DR a reality
Low-cost options for DR site-to-site replication tools