What you will learn from this tip: This tip discusses some of the elements that should be considered when selecting the location of a remote or alternate recovery site.
It has become a widely accepted fact that an increasing number of organizations have to manage and protect massive amounts of data on a daily basis. It is also a known fact that, while most IT departments are comfortable with the success rate of their tape-based backups, most know that they likely could not restore from backup tapes within the business' recovery time objective (RTO). This has lead to an increase in popularity of disk-based backups and remote data replication products of all sizes.
Many companies have implemented or at least considered the implementation of a data replication product at a remote site. It has become clear to many that the only way they could ever hope to regain access to their business' critical data within established RTO is through remote data replication. However, what is the actual definition of "remote" in a recovery situation, and how far away is far enough?
The same questions resurface on a regular basis: What is the industry norm for the distance between production and alternate or DR sites? Are there legal requirements? What are other companies doing?
The short answer is that there are no standards or specific rules at this point in time -- even the SEC recognized that imposing a "one size fits all" ruling wouldn't be a simple task (Click here for more information). The reasoning is that threats and probabilities vary widely between geographies. While 10 miles between two sites may be suitable in the event of a tornado or a fire, this distance would be insufficient in areas prone to hurricanes, for example, like the Gulf Coast after Hurricane Katrina.
Numerous factors must be considered when establishing the distance that should separate two sites. They can be summarized as follows:
- Threats and probabilities
An alternate location is chosen based on which type of threats your environment is exposed to. For instance, if fire presents the highest risk, a remote site across a campus may be adequate. Conversely, an increased risk of a widespread disaster, such as a hurricane, will dictate much greater distances between sites. Statistical data can provide guidance in establishing probability of occurrence.
- Risk tolerance
Potential risks can also include widespread power grid failure (2003 blackout), wide-area weather phenomenon (1998 ice storm), etc. An organization must assess risk based on existing threats and vulnerabilities measured against probabilities. Because risk can be mitigated but rarely completely avoided, a decision must be made as to what represents a risk that is tolerable to the business.
- Potential losses vs. cost of resilience
This is probably the most important factor to consider. The cost of establishing and maintaining a remote site should never exceed the potential losses incurred by the absence of that remote site.
The RTO will dictate whether a remote or alternate recovery site is required for a specific service. An RTO will also help determine whether failover to a replicated environment (hot site) is required or space to accommodate recovery is sufficient (cold site).
- Remote site accessibility
Although listed last on the list of criteria, it should be on every disaster recovery planner's mind when selecting a suitable location for a remote or alternate recovery site. Serious consideration must be given to the way applications and data at the remote site will be accessed by business units and end users. Lack of network bandwidth or accessibility may render the remote storage arrays as useful as space heaters in an air-conditioned room.
Regular risk assessment and business impact analyses will ease the task of measuring potential losses. These two business continuity elements should also be part of any sound enterprise risk management program.
About the author: Pierre Dorion is a certified business continuity professional for Mainland Information Systems Inc.