"Recovery of Critical Information From Your Computer System"
J. Patrick Guider, Jr.
President, TOPAZ Technologies, Inc.Content: Disaster Categories - Formulate A Plan - Backup Tools
Disaster Recovery
Computers have revolutionized our lives. They are in our homes, cars, pockets. Computers control
our money, our data and in many cases, our work is almost totally dependent on the performance of
our automated systems.
What do we do when something goes just a little bit wrong…Disaster Categories
The first hurdle to overcome when planning for disaster recovery is to realize that the seemingly large variety of possible disasters can actually be reduced to a manageable number. In point of fact, all disasters can be grouped into one or more of only three categories. These are:
• loss of information
• loss of access
• loss of personnel• Loss of Information
Loss of information may be caused by several factors:
- Accidental Deletion.
- Deliberate Deletion, e.g., disgruntled employee.
- File Corruption due to improper shutdown or disk damage.
- Hacking/Viruses.
• Loss Of Access
• Problems Leading To Loss Of Access
• Forgotten/Lost Passwords
• File permissions improperly set or corrupted.
• Hacking – Denial of Service (DoS) attacks
• File permissions – users set to Read Only, nobody knows a Write Access password.
• DoS Attack – Simply when someone does something to stop a server from serving data
(Syn Flood, Ping of Death, etc.)
• Loss Of Personnel
Cantor Fitzgerald
More than 700 employees lost in World Trade Center attack• Key personnel unavailable due to sickness/injury/death.
• Terminated employees leave without revealing passwords to critical systems.Akamai Corporation kept running “Still, Akamai's technology carried much of the redistributed load when
the loss of the World Trade Center echoed through the Internet. Across its native internet habitat, the Linux
community has taken to heart the business half of the Slashdot slogan: Stuff that matters. Quietly, with no
publicity, enormous rebuilding work is already being done. “
• Risk Analysis – You must know what potential problems you could face.
• Prioritize The Operations – What operations are mission critical to keep the organization running?
Think about payroll and other financial requirements.
• Inventory – What are the organizational assets? What must we safeguard/replace to survive?Databases
• Relational database system files should be backed up on a regular basis.
• Ensure that mission critical information is maintained in enterprise level RDBMS
(Relational Database Management Systems) rather than “desktop databases”
(MS Access, Paradox, etc.) due to their relative instability and lack of file recovery tools.Best Practices
• TRAIN, TRAIN, TRAIN– Ensure all personnel are well versed in their rolls in a disaster recovery operation.
• Plan, schedule and conduct Disaster Recovery Drills using your plan. Assess weaknesses and correct as required.
Safeguard Your Data – Backup Tools
• Murphy’s Law applies to mission critical computer data. “If it can be lost, damaged or
destroyed…it will be.”
• Tools exist to keep the data of an organization secure:• Tape Drives
• Magnetic Optical Drives
• Removable Disks
• Online Backup
• RAID (Redundant Array of Inexpensive Disks)
Tape Drives
• Tape drives are the most commonly used backup device
• Magnetic backup tapes are inexpensive
• Backup tapes are the slowest method of backup
• Tapes wear out over timeMagnetic Optical Drives
• Use 3.5” optical disks that look similar to floppy drives.
• Disks are expensive.
• Disks do not wear out as fast as tapes.
• Can be used in tape “jukeboxes” that allow for the use of multiple tapes.Removable Disks
• Good backup speed.
• Media is readily available.
• Tapes are expensive.Online Backup
• Throughput depends on the type infrastructure you have in place at your facility
• Offsite storage is attained automatically
• Cost depends on the amount of data backed upRAID (Redundant Array of Inexpensive Disks)
A group of hard drives under the control of array management software that work together to improve performance and decrease the odds of losing data due to mechanical or electronic failure by using such techniques as data striping. Striping is a method of writing sectors of drives across multiple hard disks.
The most common RAID levels are RAID 1 and RAID 5. RAID 1 involves one hard drive being copied (mirrored) to a second drive. RAID 5 uses a sector based striping scheme.
• All data is backed up to magnetic storage tape on a scheduled basis.
• Backup tapes are stored offsite at a secure location.
• Backup tapes are tested AT LEAST quarterly to ensure integrity.
• All servers utilize RAID 5 redundancy.
• Antivirus software definitions are updated at least once a week on both servers and workstations.
• Process and systems are audited by the Quality Assurance Department
SummaryThere is no excuse for data loss.
Why?
• The technology exists today that will prevent it
• The resources to put in place a recovery system are far less expensive than loss of mission critical data.
Planning for Disaster Recovery can cost time and money…
...but how much is your data worth?
Real Life Disasters - Disaster Planning - Facility Planning - Assets Protection - Web Links - Home
copyright 2003
