Sys Admin and Redundancy And Redundancy By Brandon

Sys Admin and Redundancy And Redundancy -By Brandon Fuller & Christopher Gossett

Agenda � Basics � Pros and Cons � Redundancy in terms of � Hardware � Software � Personnel � Finer points � Conclusion

Basics �Redundancy: � The inclusion of duplicate systems that function in case of failure. Ultimately any part or system or machine will fail in some way. The Goal of Redundancy is to maintain Uptime despite failing machines.

Benefits of Redundancy �Decreases the chance that a system will fail when any one redundant part fails. �Increases Uptime �Better allows a system to endure disaster, especially with remote redundancies.

Drawbacks of Redundancy �Increased cost. �Increased complexity �Increased Storage space

Redundancy in Hardware �Most familiar among most people. �Additional hardware to function when main part fails. �Extends beyond the direct computer, also includes the cooling systems, power supply, and storage. �Cost of redundant systems outside of computer can outweigh cost of computer.

Redundancy in Software �N-version programming �Functionally equivalent programs are created by different groups with the same specs. �In theory very low chance of all versions failing at the same time. �Redundancy in software can be used as a cost alternative to certain hardware redundancies (often at a cost of speed)

Redundancy in Personnel �In order to ensure that, if you are on a team of sysadmins, that you are not on call at all hours. �Have other employees who also know how to fix and respond to certain issues. �Can be problematic if they become too redundant. Job security might suggest you being a sole point of dependence.

Redundancy vs. Backup �Not the same thing. �Backups don’t maintain uptime. �Backups are also often periodic and may mean data loss. �Backups can fix mistakes. �Backups can keep a record.

Points of Failure �A point of failure is any place in the system that can break and cause failure. �Points can range from as small as a wire or chip, to as large as the entire building. �Consider redundancy for multiple points. �Remember: A thousand redundancies of point A do nothing if point B breaks with no redundancies. �Consider remote locations if data is important enough. � 100 miles for large scale insurance reasons.

Passive vs Active Redundancy �Passive: �Use extra parts to reduce impact of part failures. �Failures normally cause performance decline however. �Example: Humans �Active: �Removes performance decline by monitoring performance of each part. �When a faulty part is detected activity is rerouted to other parts without performance drop. �Example: Power Wires

Conclusion �Redundancy helps to ensure uptime. �Redundancy can be expensive. �Important consideration when planning a system. �Keeping up redundant parts prevents system failure. �Be careful of single points of failure. �Redundancy helps to ensure uptime.