MicroNet Technology RAIDBank4 manual Glossary

D-Glossary

SCSI acronym for “Small Computer System Interface”, SCSI is a set of standards for physically connecting and transferring data between computers and peripheral devices. The SCSI standards define commands, protocols, and electrical and optical interfaces. SCSI is most commonly used for hard disks and tape drives, but it can connect a wide range of other devices, including scanners and CD drives. Most modern storage protocols are based on the SCSI Block Protocol.

Secondary Storage Mass storage devices such as hard disks, magneto-optical disks, floppy disks and tapes are frequently referred to as secondary storage.

Stripe A stripe is a logical space that spans across multiple hard disks with each constituent hard disk contributing equal strips (or chunks) of space to the stripe. In the figure below, strips 1, 2, and 3 from hard disk 1, 2, and 3 respectively comprise a (purple colored) stripe. Synonym: major stripe

Stripe Set A stripe set is a set of stripes that spans across multiple hard disks. In the figure below, the displayed stripe set has 4 stripes, with strip number 1 comprised of the purple strips 1A, 1B and 1C. Stripe number 2 is comprised of the green strips 2A, 2B and 2C etc.

Stripe Size This is the size of the strips that constitute each stripe. This term is a misnomer

– though prevalent – since it should appropriately be called strip size or chunk size.

TCP/IP This is an acronym for “Transmission Control Protocol/Internet Protocol”. It is comprised of two parts TCP and IP. The former, i.e., TCP is a peer-to-peer connection oriented protocol that guarantees the delivery of data packets in the correct sequence between two peers. The latter, i.e., IP is the protocol that defines and governs addressing, fragmentation, reassembly and time-to-live parameters for packets.

Volume Set A volume set is a concatenation of storage elements that may be RAID arrays, JBODs, or simply areas of disks that are not part of RAID arrays.

Write-back Cache a caching scheme that acknowledges the write request as complete before data is written to the final storage location. This methodology can improve the efficiency of write operations under favorable circumstances, but is at risk of data incoherencies in a system that is not protected from power fluctuations or failures.

Write-through Cache When a cache is operating in write-through mode, data written into the cache is also written to the destination secondary storage devices. Essentially write completion does not occur until the data is written to secondary storage. Thus the contents of the cache and the secondary storage are always consistent. The advantage is that the possibility of data corruption is greatly reduced. The disadvantage is that write-through operations are more time consuming

XOR Function All RAID arrays (with the exception of RAID 0, RAID 1 and RAID 10) require parity to be calculated and written to the array in conjunction with data. Typically the parity is a simple XOR on the bytes comprising a stripe. This is a computationally intensive operation that many modern RAID controllers perform using a dedicated ASIC often referred to as a XOR-engine.

RAIDBank4 Owner’s Manual 63