Week 14: Clear

In this week's lesson, I learned about the multitude of ways in which a business or enterprise can store its data. Options include HSM (where data is migrated to cheaper storage if it is not used frequently in order to conserve valuable expensive storage for the data that is used frequently), SAN (where a network is designed specifically for sharing storage resources among a set of servers), Virtualization (a technique for running multiple "virtual computers" on a single physical host computer system, or through a Thin Client (small inexpensive--compared to a personal computer-- terminal-like computer that mostly relies on a server for data storage and processing.  This week began with sort of a throw back to the start of the course with the types of storage methods.

RAID 1 (Mirroring)--which involes a pair of drives, where each block of data is written to both drives. If a single drive was to fail, no data would be lost and the system would still run at full speed. This creates redudancy, which is great if you want to protect your data from drive failures. RAID 0 (Striping)--which involves two or more drives, where blocks of data spread across all drives equally. This creates a fast read and write speed due to multiple disks simultaneously working. There is no redundancy which means that if one drive fails...all data is lost. It is statistically less reliable than a single disk drive in that the more drives there are, the greater the total chance of there being a failure. Raid 5 involves three or more drives, where data blocks spread across all drives except one. This exception drive is used to store "parity" data--this is where data in the form of ones and zeros are added up to formulate an even or odd number..if even, the data is stored as "0" on the parity drive...if odd, the data is stored as "1" on the parity drive. This creates fast reads similar to striping, with moderate write speed due to the computations involved with storing data as parity. If a drive fails, no data is lost but the system will run at a reduced speed because the parity drive is being used to compute the data lost from the missing drive. There is even a "hot spare" which is an extra hard disk that remains idle only until a drive fails...then it will automatically begins to mirror the infomation of the lost drive. Finally, there is Raid 6 which is any RAID that can continue to run if there were to be two drive failures at the same time. It's performance is similar to RAID 5 and it protects agains read failure during rebuild of big disks which could result in data lost.