Planning for Disaster Recovery

An Overview Of The Issues

Business disasters that might affect your information-technology infrastructure can come in many different forms and degrees.  The possibilities range from corruption of a single important file, up to complete destruction of the building that houses your office.  The September 11 destruction of the World Trade Center is heavy on our minds and at the forefront of our thoughts as we address this issue of disaster preparedness.  But potentially more devastating still is the prospect of a major earthquake such as the San Francisco Quake of 1906, which caused destruction on a regional scale.

Bay Office Computing has developed a five-tier disaster scale to aid in discussing specific redresses for different degrees and types of disasters.  These five disaster levels in order of increasing severity are 

Each are discussed below.

File Level

Occurrence

A file level disaster can occur in several ways, but there are two scenarios that cause the vast majority of these problems in a typical office setting.  One is accidental deletion, and the other is file corruption.  Often awareness of this type of disaster occurs immediately. A more insidious problem is when the deletion happens but is not noticed for a long period of time, perhaps months. 

Protective Measures

Strategies for protecting and restoring individual files centers on backup strategies and technologies. 

Workstation Level

Occurrence

Workstation-related disasters can result from hardware failures, or from Operating system failures.  In either case, the workstation is rendered unusable, and employees lose productivity, or worse, valuable work.

Protective Measures

There are a number of ways to protect against lost productivity and data due to workstation failures.  Hardware failures are the most difficult to protect against.  Overheating and dust are particularly damaging.  

Operating system failures usually result from improper user interventions, such as loading software and adjusting system settings.  User training, system policies, and file-saving protocols all come into play.  In addition to these measures, there are design and configuration options that can reduce the impact of a workstation failure to a minor inconvenience.

Server/Network Level

Occurrence

Server and Network disasters occur due mostly to hardware and operating system failures.  Additionally, networks can go down from failures in the cabling plant, which consists of hubs and switches that link all of the network cables.

Protective Measures

Fault Tolerant Design

There is a relatively short list of essential fault-tolerant design criteria which servers should incorporate.  Servers should also be made with quality components from a reputable manufacturer who also offers good support services.  For added redundancy, it is always good to have multiple servers with backup and redundancy accounted for in their functional design.  Additionally, provisions should be made for backup power and automatically triggered server shutdowns before powering-off.  Additionally, email programs can be configured to provide local backups of server-based mailboxes.

Data Backup

While data backup was mentioned at earlier, it is important enough to be revisited here with some additional information:  Tapes should be secured onsite against fire, theft, and other conceivable damages.  Tapes should also be kept offsite.  Rotation cycles to accomplish both are standard BOC procedure.

Office Level

Occurrence

This type of disaster occurs when there is a building fire, terrorist event, or natural disaster.  In a case where an office and all the networking equipment in the office are destroyed without warning or notice, office functionality can still be restored with proper planning.

Protective Measures

Off-Site Data

Of course, the cornerstone for restoring network functionality in the event that the office is physically destroyed is having your data protected offsite.  And the centerpiece of offsite data in a small office is the tape backup.  This fact again underscores the importance of scrupulously maintaining the tape backup system

Other options included offsite sychronization and maintaining redundant offsite hardware.   BOC has numerous detailed options for these scenarios including options for restoring functionality in 24 hours even if the office is completely destroyed.

Local Area Level

In the event that a region wide disaster was to occur, such as an earthquake, major disruptions to all businesses are likely.  It is conceivable that such an event would also disrupt phone lines, data lines and Internet access including services provisioned by ISPs.  Thus, in the event of a regional disaster, a longer time frame to restore functionality is more likely the best-case scenario. 

Conclusions

There are many levels of disaster with different levels of severity.  But recovery from all of them hinges on one common and therefore critical element, a good tape backup.   The firms tape backup strategy should therefore be scrutinized and gaps filled.  A successful strategy includes not only after-the-fact actions, but fundamental elements of designing and maintaining your network and workstations.