DISASTER TOLERANCE

Preparedness, Protection For Your Business

By SAM BISHOP & RON KRAMER

For businesses relying on mission-critical computing systems, downtime isn’t an option. For emergency-response organizations, lost time can be life threatening. In the business world, downtime and lost productivity mean lost revenue, an unacceptable proposition in any economy. The best way to reduce downtime is to prevent it in the first place by deploying continuous computing technology.
The goal of fault tolerant solutions is to provide nonstop, 24/7 computing, even in the event of a component failure or software crash. Until recently, fault tolerance has been predominantly Unix-based and these costly systems can be justified only in the high-end enterprise market. Small to mid-size businesses operating in a Windows/Intel or Solaris environment – manufacturers, call centers, utilities, and e-commerce companies – are looking for assured availability to protect their data and their investment.
In the past, these types of businesses have had little choice but to rely on backup systems and failover tools in order to recover from unplanned outages. Disaster recovery means a crash has already occurred. Today, affordable fault tolerant solutions can be deployed to prevent outages from occurring in the first place – to completely eliminate recovery time. This concept is known as “disaster tolerance.”

Cost Of Downtime
Most business computer systems operate at 99 percent uptime or less. This equates to 10 business days a year of downtime. Mission-critical systems requiring 99.999 percent (five 9s) availability means less than six minutes of downtime in a year’s time. Six 9s of availability means less than one minute a year of downtime. The types of business commonly associated with this level of uptime include E911, toxic chemical control systems, brokerages and credit card processing.
However, the use of the Internet in our 24-hour global economy has changed customer expectations. In some cases, downtime can threaten the survival of a business. There are many studies on the actual costs of downtime but business executives know lost productivity and lost transactions can have a devastating effect on customer relations. Can you recover from these losses?

Assessing Your Needs
How much data can you afford to lose during a failover or recovery? How long can you afford to wait before you are operating fully? What is the impact of downtime on business performance and the relationship with your customers? What is your IT staff’s skill set?
The answers to these questions will help determine whether a low-cost disaster recovery plan such as tape backup, or investing in a fault tolerant technology, is the right solution. A critical assessment of your company’s data and application needs, as well as the impact of downtime on productivity and customer satisfaction, will aid in the development and implementation of a disaster tolerance plan.

Deploying The Right Solution
The first step is a complete analysis of your network including the systems that collect, store and process your data. The analysis should also probe security, performance bottlenecks that can cause failures, and gaps in your backup and failover systems including a power assessment. Next, calculate the cost of downtime.
The relationship between data and downtime can be stated as an equation. As the data approaches mission-critical stage, the effect of its loss becomes greater. The cost of downtime is directly proportional to the effect of the loss of that data.
There are five categories of a disaster tolerance plan: security tolerance, management tolerance recovery tolerance, geo tolerance, and fault tolerance. The equation and these five T’s make up our “disaster tolerance model.” Each tolerant solution is dependent on where your business fits into the equation. At what point does the data loss begin to impact your operation? Choose the right tools to fit the level of tolerance in each category.
Security tolerant solutions incorporate the implementation of a company policy as well as hardware, software and network topology. Management tolerance is the ability to perform an ongoing assessment of the performance of your network by deploying a combination of in-house and remote management software tools.
One recovery tolerant solution commonly deployed is tape backup, where critical data is routinely downloaded to tapes. These tools do not protect applications however, and leaves a gap in data between the time of the last backup and when the system is restored. Additionally, its weakness is exposed upon answering the following: when was the last time you verified the backup with a restoration test?
Tape backup can be used in combination with other tools as a means of storing and archiving data. A more advanced way to protect data is offsite replication over a network connection. This geo tolerant solution is vulnerable to problems in the network.
The highest level of geo tolerance can be achieved with a fault tolerant solution in two different locations. If a disaster destroys one system, the other continues to operate with no loss of data and no loss of transactions.
Fault tolerant solutions work in settings where ease of operation or even hands-off operation is required. Fault tolerance is the right choice for businesses located outside of major metropolitan areas where knowledgeable and trained technicians are difficult to find and more difficult to afford.
Database failures, hardware failures and operating system failures can wreak havoc on your business. Be prepared with assessment and planning. Be protected with the right disaster tolerant solution. It is your fiduciary responsibility to do so.

Sam Bishop heads up business development at All Computer Solutions, Inc., a business-to-business technology solutions provider with offices in Petaluma, Calif., and Portland, Maine.

Ron Kramer is the firm’s vice president and system architect. For more information about their disaster tolerance practice visit www.disastertolerance.com.

To comment on this article, go to 1503-16 at www.drj.com/feedback.

«BACK to the Articles Index