Protecting your Knowledge Base

Knowledge management has many different facets, all of them important in the quest for becoming a true knowledge sharing organization. But when it comes to tangible aspects of KM, technology quickly comes to the forefront of most discussions.

After having developed your overall strategy, implemented the procedures, and established the technology infrastructure to support your KM initiative, how do you protect the valuable resources that are now being captured and stored? Are you sure that the explicit knowledge that resides in the files servers, application servers, web servers, e-mail server and database servers is secure from potential disaster?

With the seemingly increasing number of weather-related disasters hitting various parts of the world nowadays, backing up your knowledge warehouse and protecting it is becoming a popular topic of discussion.

So what trends are evolving in regards to storage and backup?

With the rapid deployment of numerous application servers to meet end user and corporate requirements, managing them efficiently - and the data they hold - can be a real nightmare. As a response to this, "dedicated storage networks have become an emerging trend," so says Jonathan Greene, Computer Associates' Director of Storage Product Marketing. In addition to ease of administration, centralized storage users can see an increase in overall performance and availability.

Is it a SAN, LAN or WAN?

Although centralized storage has been around for quite some time with mainframes, bringing it to the enterprise network is a concept just beginning to take off. While many are familiar with Local Area Networks (LANs) and Wide Area Networks (WANs), Storage Area Networks or SANs are a relatively new area in enterprise data access and backup.

A SAN is basically a dedicated sub network consolidating all storage and backup devices, that is connected via a fibre channel to your corporate network. Instead of having multiple storage devices (hard drives disk arrays, CD-ROM towers, etc.), each attached to individual servers across the organization's campus, SANs bring all these devices into one centrally managed location.

Not only are SANs easier to administer, they make more efficient use of available storage space since multiple applications can reside on a single device. Computer Associates' Greene believes that "storage can no longer be managed in isolation or thought of as attached to a server."

Window is getting smaller

Chris Ilg, Senior Product Manager at Exabyte, a manufacturer of tape backup solutions, says speed of the backup should be the primary issue for companies. "Customers are demanding faster backup methods more so than for larger backups. There is a premium on the available time to do a backup," says Ilg. "Before, there was an eight hour window to do a backup. Now it is four to six hours."

Ilg believes this is true because there is more end of the day processing being done and the workday is being lengthened. Before, the end of the day meant 5:00 or 5:30. Now it is 7:00 or later. When the day ends at 7:00, and after the time needed for typical end of the day processing is done, the remaining backup time window has shrunk dramatically. Now companies have to do the backup and have everything running by 5:00 or 6:00 am when the first staff begin to arrive for work. "The window for backup time is getting smaller and smaller when you consider the global market," says Ilg.

Rami Hyary, Regional Manager at Acuitive (www.acuitive.com) agrees with Ilg that speed is important. "Choose your backup devices according to speed, not capacity." But Hyary also says it is critical to perform periodic restores of data to test that the process works.

Have you checked your backup today?

"It is not enough just to make sure the backup system didn't report any errors," says Hyary. "When performing a backup, it is very important to remember that the purpose of doing a backup is to be able to restore the data if a problem occurs."

Hyary recalls a company where the client was doing a backup for months without checking to see if the backup was good. It was only while they were upgrading the server's operating system that it was discovered the backup tapes were no good. The backup software the company used was an older version that was just verifying the file header to check if the backup was successful and not the file itself.

As Hyary quickly found out when the call for help to his company was made, a hardware failure had caused the transposing of data bits, but was still sent the same amount of data. As a result, the file size did not change so the backup software did not detect a problem. The end result was that the accounting department had to rekey in three months of financial data.

Hyary suggests to "regularly perform a routine of creating a file with known data in it. Then let your regular nightly backup procedure take place. The next morning, delete the file and wait one day or one week, then restore the file from tape. Did it work? Make sure you know from time of failure, how long it will take you to make the data available to users." Is this time acceptable to your users?

Where to begin?

There are many levels of protection and steps that are needed to safeguard your valuable knowledge warehouse. CA's Greene suggests "starting with the mapping of critical resources" (sounds like KM, huh?) and ask yourself the following questions:

What amount of downtime is acceptable? If your servers go down during the day, how much data can you afford to lose (since your last backup)? Can you reconstruct the lost information? How long would it take? What personnel resources are available in the remote sites? By answering the above questions, you are taking a big step in developing a solid backup plan.

Other Methods

Besides backing up your data to another media such as tape, there are other steps to take to insure maximum up time of your data.

Although not a new storage trend, RAID (Redundant Array of Inexpensive Disks) has become a popular protection method for individual hard drive failure. Instead of storing all you data on a single hard drive, a RAID configuration can store the same data in different locations on multiple hard disks (thus, redundantly). The benefits are added security if a single hard drive fails and increased performance since reads and writes of data can occur simultaneously on multiple drives.

There are ten configuration levels of RAID, each offering a different level of redundancy and performance. Typically, if one drive fails (such as in a RAID 5 setup), the users won't notice a thing. The system keeps running as if nothing happened and the network administrator gets a notice that says there was a drive failure, which can be replaced after hours.

Although RAID 5 is designed for individual hard drive failure, what happens if there is another failure in the system, such as the network card or motherboard? How do you prevent this?

Enter the hot server. A hot server takes over from your primary server in case of a hardware failure within the primary server. Similar to mirroring one hard drive to another, hot servers mirror the compete server, including data written to the drive in sub one second intervals, to a backup server that is ready to take over if needed. So, if the primary server goes down, the secondary server takes over and you are back in business in seconds instead of hours or days. Companies that provide these types of solutions include Vinca (acquired by ?) and Octopus.

Another option is to take the storage services completely off-site. Even with a hot server, the secondary server may not help if damage was done within your server room by a disaster such as a storm. To aid in these situations, companies either can create a separate location with duplicate data that can take over when needed or turn to a third party that offers these services.

Back to Basics - Don't forget the paper!

Giff Salisbury, CEO of Commercial Archives (Buffalo, NY), has developed a disaster recovery program that he not only markets, but he also utilizes in his own business. In Salisbury's program, he discusses his 20-step program for clients to develop their own disaster recovery plans or, if they choose, companies can use his services to assist in the designing of their plan. Salisbury points out an important thing to keep in mind. "Many times we think of only the computers within a company, but what about the paper documents?" Could your business recover if you lost all your hard copy documents in a disaster? How long would it take?

With the vast amount of information that can be captured and stored today, companies are becoming more and more dependent on these repositories. These massive databases and storage devices are growing exponentially and so should our need to provide a reliable backup and recovery plan.

Without the ability to replace the information we use to create knowledge, where would we be? Are you providing a safe haven from the next storm for your most valuable asset - knowledge?

Stephen Goodfellow is President of Access Systems (www.AccessKM.com), an independent consultancy in Knowledge Management and technology assessment. Steve can be reached at (315) 682-1188 or via e-mail

KM World Magazine

Send mail to Webmaster with questions or comments about this web site.
Copyright © 2005-2008 Access Systems, Inc.
Last modified: September 19, 2008