Storage

Data is written/read across two or more drives simultaneously, effectively allowing it to be accessed much more quickly. It allows all the hard drives in the RAID array to appear as a single device with the total combined storage of all the drives. E.g. two hard drive of 256GB and 512GB, would appear as a single 768GB drive. This form of RAID is sometimes called striping and does not provide any extra redundancy than storing your data to a single hard drive. If one drive fails and is then replaced, the other can not rebuild the whole data file.

Two or more disks are used to store the data as exact images of the other. As data is being written or read from one disk the RAID controller “mirrors” this data to the other drive/s. This provides a high level of redundancy as if one drive fails the other has a full working set of data. However, it means that you would only be able to have as much storage space available as the size of the smallest disk in the RAID array. E.g. two hard drive of 256GB and 512GB, would appear as a single 256GB drive.

A combination of RAID 0 and RAID 1, data is distributed across multiple drives without parity, and then the entire array is mirrored. Although this delivers good performance, the drive storage overhead is 50% because you are mirroring the data.

Data is striped at a block level across several drives, with parity stored on one drive. The parity information allows recovery from the failure of a single drive. The performance of RAID 4 is very good for reads. Writes, however, require that parity data be upgraded each time. This slows down random writes in particular, though large write or sequential writes are fairly fast.

Employs data striping and parity across all drives in the array creating better performance and security. Since parity information is striped across all drives, lost data can be retrieved and rebuilt from the parity.

All-flash storage array is running SAN or NAS protocols on a set of controllers and is best when optimized software is engaged to take advantage of all enterprise SSD drives allowing quick IOPs and throughout.  All-flash systems will not run spinning hard drives and must be all SSDs.   

A method used for storing information on multiple devices. In storage terms ‘array’ is commonly a collection of hard disk drives in a server arranged in a particular way, that stores the same defined data as each other but can have different values. 

After data has been written to the primary storage site, new data can be written to that site, without having to wait for the secondary (remote) storage site to also finish writing data. Asynchronous Replication does not have the latency impact that synchronous replication does, but has the disadvantage of incurring data loss, should the primary site fail before the data has been written to the secondary site.

New technology first introduced by Compellent around 2007.  EMC and 3PAR also came on strong around 2008 with their solutions.  Auto tiering is most commonly referred to; data residing on two or three classes (performance) of storage.  These would most commonly be today; SSD, 10k and 7200RPM drives. Autotirering resides on the controllers and as data is utilized by the user, data moves to the appropriate performance storage.  For example, high-speed data being written to the storage can write to SSD then move to lower-cost 10k drives. Once the data becomes less required, the data can move to lower-cost near-line 7200RPM archive drives.  

Backup is the process of replicating your vital data onto a secondary storage device or off-site storage, for the purpose of recovery incase the original data is accidentally erased, damaged, or destroyed.