High Availability: Is it Sufficient for Disaster Recovery and Business Continuity?

Abstract
Business and information technology (IT) executives are increasingly accountable for business continuity planning because of a confluence of forces: regulatory mandates, rising awareness of best practices and high-profile disasters that have drawn attention to risk. While concepts such as “high availability,” “continuous availability,” “disaster recovery” and “business continuity” have been around for years, today they merit a more discerning look.
While the intent of high availability is similar to continuous availability, and disaster recovery is closely related to business continuity, there are important differences in their scope, methodologies and outcomes. And although these terms are often used interchangeably, it is important to understand what they truly mean in terms of your business’ IT planning and execution model.
This paper will examine how availability management, disaster recovery and business continuity support one another _ and offer insights to help you succeed at all of them.

Higher Stakes for IT Availability
IT is so tightly woven into the fabric of business that commerce on any significant scale doesn’t take place without it. The same is true for public safety and government operations. Minute by minute, business and organizational processes depend upon a multitude of computing applications and their associated data _ for VoIP communications, supply-chain RFID, financial transactions, POS integration, emergency services dispatch, electronic medical records and much more.
For mission-critical IT, service degradation, service outages and data loss pose serious consequences. Maintaining service levels is anything but simple under normal conditions. Knowledgeable technical professionals, skillful management and prudent technology choices are all required. Even then, numerous threats can bring disruption in the blink of an eye: human error, computer viruses, malicious intrusions, natural and man-made disasters.
The costs and consequences of an even an hour of downtime can be staggering. Money, property, competitiveness and reputations hang in the balance. In public safety and healthcare, human life can be in jeopardy.
However, it’s imperative to understand the monetary and non-monetary impacts of service degradation, service outages and data loss in your environment. Consider different scenarios. For example, the costs and consequences will be very different during periods of peak and low activity. Your business may in addition be subject to regulations that mandate complete, auditable data _ making any data loss a concern.
Also don’t be too quick to dismiss an application as non-mission critical. Think of: the e-mail server during the last days of the quarter when your sales force is closing orders; the value of the business transactions that flow through your IT infrastructure during your busiest season.
Evolving regulations and best practices, the latter including ITIL® and COBIT, are also elevating the priority of IT service and data availability. Yet a gap widely persists between many organizations’ expectations, and the knowledge of what is actually necessary to keep essential business processes moving at maximum velocity.

High or Continuous Availability?
When your business relies upon mission-critical applications, the server is a logical place to enter the discussion of availability. Let’s first establish the meaning of a few main terms:

• Availability: describes the condition of IT system(s), component(s) and/or networks functioning as expected when needed; used interchangeably with uptime.

• High availability (HA): describes the availability level of IT systems designed to minimize unplanned downtime through quick recovery from failure.
• Continuous availability (CA): describes the availability level provided by IT systems designed for uninterrupted operation.
• Availability, or uptime, is customarily expressed as a percentage. Usually that percentage reflects the effect of unplanned downtime only. Downtime for scheduled maintenance, such as software upgrades, is excluded. That’s meaningful because a system may be built to minimize planned downtime _ or not.
• HA systems vary in the levels of availability they promise, generally between 99.9%-99.99%.
• CA systems are engineered to be available 99.999% of the time or more. The differences in availability can be explained by the systems’ architectures and how they are supported.
• In the HA category, a major contender is clustering. A cluster links two or more computer servers with software programs and physical connections so that when a failure occurs on one, its workload fails over to a server that is still operating.
• Users are interrupted during a server outage but may be able to log back on quickly.