The Disaster Recovery (DR) stakes are higher than ever. No longer is being able to recover “eventually” an acceptable strategy. Users and application owners at organizations of all sizes are expecting to be back online within hours, no matter how significant the disaster. The problem is IT budgets are not increasing at the same rate as DR expectations. Lack of DR budget is forcing IT to create more innovative disaster recovery plans. One option under consideration is “the cloud”. Cloud backup and DR promise to lower costs and decrease recovery times, but using the cloud comes with its own set of problems. IT planners need to factor in these potential problems before making the leap to Cloud DR.
While there is a lot of information, mostly found on vendor pages, on how wonderful cloud backup and DR is, there is not a lot of information on its potential downsides. In this column we will introduce some of those challenges. For even more information watch our on demand webinar “Cloud DR, is it worth the cost?”.
The Problems With Cloud DR
The most pressing concern that IT planners express when beginning the move to the cloud is security. Security though is a manageable concern. The first step to feeling more secure about cloud-based data is to understand how these solutions provide encryption. It is important to look for a solution that gives your organization control over the encryption keys. The goal is to have the data locked from the time it leaves your data center and while it is at rest in theirs.
The second concern is a series of concerns that can be best summed up as issues with latency. The problem is that you are squeezing a lot of data down a comparatively small pipe. On the backup side, techniques like deduplication and compression help considerably, making the sending of data to the cloud a reality. On the flip side is recovery and deduplication won’t help here. All the data comes back. The solution is to make sure your cloud provider can deliver data to you by shipping disk or tapes to your facility. Slow replication back is not going to be acceptable. Another workaround to the recovery problem is Disaster Recovery as a Service (DRaaS) which instantiates the organization’s applications as VMs in the provider’s cloud. The problem is that DRaaS has its problems.
DRaaS eliminates the bandwidth problem that cloud providers face regarding disaster recovery, but it creates a new set of problems that IT planners need to work their way through. The first is how long will it take for the organization’s applications to be online after a declaration of disaster? Some providers promise “instant” while others a couple of hours, and many provide no guarantees. Once the organization’s applications are running in the provider’s cloud, the other big issue is how does the organization get out? Eventually, all that data has to be recovered to the production site. DRaaS makes the final recovery more complicated since data is changing in the provider’s cloud while it is in transit to your data center.
Finally, there is cost. As Storage Switzerland has pointed out numerous times, while the initial payment for cloud storage is very attractive, the 100th is not. Cloud backup, if not managed, can get out of hand quickly. Storing year’s worth of backup data in the cloud is very inefficient from a cost perspective. Organizations retaining large datasets for more than a few years will want to explore an aggressive archive strategy that keeps archive data out of the cloud, leveraging the cloud for recovery of active data only.
The cloud can be effectively leveraged for DR, but like any other IT solution it needs to be managed. The good news is the IT professional is still a critical part of the strategy. Storage Switzerland suggests a blended strategy that takes advantage of the strengths of both on-premises and the cloud while minimizing the weaknesses of each. In this week’s webinar, we’ll detail the pros and cons of cloud DR as well as articulate our blended cloud/on-premises strategy.