TECHNOLOGY SERVICE NEWS

QUARTERLY REPORT October 2005, Vol. 2, Issue 4 ©2005

Plan Today for Disaster ‘Recovery’ Tomorrow By Thomas McDonald, NSI Chief Operating Officer Each year many data centers in Connecticut will experience significant service interruptions. These businesses will be threatened by a wide range of problems - from user errors and viruses, to hardware failures and natural disasters. Business Continuity is now at the forefront of IT strategies, and for good reason. This article will cover the important elements of an effective disaster recovery strategy. The first step is to create a Disaster Recovery Team (DRT) made up of department heads and other key personnel within your organization. The role of a DRT is to share information, hypothesize worst-case scenarios and circumstances, and then the best possible outcomes. The DRT’s work should lead to a Disaster Recovery Plan that can be implemented at the drop of a hat to minimize downtime. This may seem like a lot of work, and it is, but the reward is ensuring ongoing operations during a disaster, and perhaps even your organization’s survival. According to research conducted by PriceWaterhouseCoopers, 90-percent of organizations without a pre-existing disaster recovery plan go out of business within 18-months of an adverse event. Once a plan is in place, keep in mind that it is a dynamic resource because of the changing nature of our world. For example, you may have added more severs to your IT infrastructure, industry standards may have changed, or new government regulations enacted. We recommend that the DRT meet every three months to review, update, and test the disaster recovery plan. Different disaster scenarios should be assigned to different personnel. One such common scenario might be recovery from a fire. More than one person within the organization should be trained in fire response – location and use of fire extinguishers, evacuation plans, hierarchy of data protection, notifications, etc. – in the event other trained personnel are off-site. Don’t just have backups for hardware and software; have backups for your people as well, whether they are in-house, outsourced, or both. Also, be prepared in advance for the large amount of documentation that is often involved with disaster recovery, such as check lists and resource tracking. In the event of a disaster, you will have to literally go from “Point A” to “Point B” documenting the steps you have taken. Another crucial aspect of disaster planning should be the consideration of some relatively new technologies that can assist in preventing disruptions, or keeping downtime to the absolute minimum. I’ll start with backup and storage. Many organizations still utilize tape backup and storage systems because the media is inexpensive. However, tapes are vulnerable to disasters, particularly to environmentally-induced failures. While tapes are sensitive to floods, fire, and the like, removable tapeless backup disks are completely sealed in stand-alone units that provide data protection, plus rapid data restoration. xSeries Tapeless Backup Tape systems are also vulnerable to human error that can lead to disasters. A case in point involved Bank of America several months ago. An estimated one million customer records on tape were “lost” during shipment to a Backup Center. We strongly urge the adoption of a tapeless backup solution as an integral part of your disaster recovery plan. While on the subject of backups, I must point out the obvious – they help most in a disaster situation if they are current backups. We recommend that backups be performed nightly and tested regularly, at least once a month. The more the better, to provide the peace of mind that your backup is ready if disaster strikes. Increasingly, another key element of disaster planning is the deployment of a “virtual IT infrastructure”, where the infrastructure is partitioned, isolating servers in secure and transportable machines. In this architecture of “continuous availability” workloads are balanced over several, often geographically distributed, platforms. Each platform is set up to have spare capacity. When a disaster causes the failure of one platform, the workload is distributed over the remaining platforms. This approach is attractive because it allows organizations to maintain operations even after a disaster has occurred. Today many organizations are also turning to specialized software for client blade servers, which have dozens of blades housed in a single cabinet, to assist with disaster recovery. With “failover” capabilities built into the hardware and software, all local (C-drive) disks can be replicated to other blades within the cabinet, or outside the cabinet. Essentially, your local drive data can easily be transported from one site to another and dozens of PCs can be up and running again within minutes. Keep in mind that the cost of including a specific application or system in your disaster recovery strategy should be weighed against the potential risk and impact of prolonged downtime due to an event. It is advisable to implement disaster recovery solutions when the costs are lower than the projected impact of the outage. IDC estimates that in disaster situations the average loss in the United States is $3-million per incident. Disaster recovery planning is serious business and it requires serious ongoing attention. It may be time consuming, unpleasant to contemplate, and costly. But when considering the options, keep in mind the old adage: “The price you pay today may be less than the price you may pay tomorrow.”