IT channel takeaway: The Backup Book author Dorian J. Cougias offers his best practices for simplifying the backup process. Keep these tips in mind when developing your own managed backup services.
With Dorian J. Cougias, CEO of Network Frontiers and co-author of The Backup Book (Schager-Vartan Books, 2003). Further insight can be found at www.unifiedcompliance.com.
Question: For a lot of companies, full backup and restoration simply is not feasible. What advice would you give them to help streamline the process?
Cougias: Backups are for the restoration of systems. How fast you need your systems back online is called your RTO, or recovery time objective. To speed the backup process and bring your systems back online as quickly as possible, you can separate your backups into recovery images, and then data backups. Recovery images are done only initially, and then after major application upgrades. They take a snapshot of the computer's volume information and save it much like a picture. Backup and restoration is very quick because the process doesn't move files around -- only blocks of data. Backups then can focus on only backing up certain directories of information that change. Once an initial "full" backup has been done, it should never be done again because all subsequent backups can backup only the changes to the data, and then create what is called a "synthetic full backup" by integrating the new data with the original data -- at the backup server -- for restoration purposes. If (or when) the computer dies, a new computer is brought in, the system's image is restored as step one and then any data is restored as step two. Neat, clean, quick.
Question: Does all data require the same level of protection? What kind of information would you recommend for, say, weekly, or even monthly, backup?
Cougias: Image your systems when you put them into service and not again until there are major changes. Your recovery point objective will tell you how much data you can lose and will set your data backup schedule. That is a mathematical calculation that only a risk manager or your corporate legal counsel can help you with. Calculate how much you can afford to lose, and then find out how long it takes to create that data. Set your backup schedule halfway between the two, and you will be fine.
Question: Tape is reliable, but slow. Disk is fast, but volatile. Is there a happy medium somewhere? Cougias: Most questions about disk versus tape are concerned with the short, medium and long term storage of data. Along with the Tooth Fairy, Information Lifecycle Management and Data Lifecycle Management are figments of the imagination that we should outgrow. Legal and regulatory compliance calls for Records Management and Records Archiving. You can be sued or put in jail for not managing records properly. No one is going to sue you for not managing data or information properly or putting data or information into some overpriced tiered architecture. The National Archive and Records Administration has set SDLT Tape as our national format for storing long term records for archival purposes. As far as media for backup, whatever you can get off site after the backup is done and be able to use for restoration in an emergency (meaning that you can test it to prove it works) should be fine.
This 3 Questions originally appeared on IT Business Edge.
This was first published in November 2006