Tape Storage
Disk Storage
Storage Networking
Home > pool > Technologies > Disk Storage
1. DASD - Magnetic Direct Access Storage Device

DASD (Direct access storage device) is the magnetic disk that data can be read and written through the random access by moving it's head on spinning disk surface. The "direct access" means that all data can be accessed directly in about the same amount of time rather than having to progress sequentially through the data.

After IBM announced RAMAC, the first generation of DASD, in 1956, DASD has continued to evolve, Removable disk pack in 1963, Floppy disk in 1970, Winchester HDD in 1973. With the RAID configurations, DASD used as a major storage media not only speed up data access, but also redundancy for reliability.

2. DLT - Optical Disk

Optical disk drives provides an effective solution to the growing need for removable high capacity storage. Optical drives are available in a number of technologies. Rewritable optical drives are based on either magneto-optical (MO) or phase-change technologies. Write-once, read-many (WORM) drives are suitable for archiving data; they can either be true ablative WORM (in which marks are ablated in the media) or firmware WORM (or MO-WORM), in which an MO disk is used but specific instructions on the media and specialized firmware in the drive prevent the possibility of rewriting. Many security-conscious customers (such as those in the financial, legal, or government fields) prefer true ablative WORM.

However, tape technologies are growing rapidly, optical disk is not used widely, but with the introduction of DVD, which can be store data up to 17GB per disk, it is searching new market opportunities.

3. RAID - Redundant Array of Independent Disk

Patterson, Gibson and Katz proposed disk array as a solution to the impending imbalance in system performance. Compare with speed of CPU and size of Memory growth, the performance of data storage access was limited by rotational speed or latency. Through the RAID configuration, they achieved improvement in terms of both performance and reliability. Today, RAID is still the best solution for providing large amounts of data storage at a reasonable cost with the added benefit of data protection



(Data Striping)

RAID 0 Implements a striped disk array, the data is broken down into blocks and each block is written to a separate disk drive. I/O performance is greatly improved by spreading the I/O load across many channels and drives, but because it is not fault-tolerant, one drive failure will result in all data in an array being lost.


(Disk Mirroring)

RAID 1 defines mirrored array. In mirrored array, a complete copy of mirrored data is written to two drives simultaneously. RAID 1 delivers twice the read transaction rate of single drive without any decreasing write performance and 100% of redundancy of data, it requires the highest disk overhead of all RAID types (100%).


(Parallel Access with Hamming ECC)

RAID 2 uses Hamming Error Correcting Code (ECC) to achieve fault tolerance and parallel access for I/O performance. It provide extremely high data transfer rates, but it require one check disk for each data disk, and read/write smaller data blocks provides unacceptable performance, so RAID 2 is almost never used.


(Parallel Access with Dedicated Parity Disk)

RAID 3 is essentially a fault tolerant version of RAID 0 that trade for capacity, for the same number of drives, to provide a high level of fault tolerance. It takes advantage of the data striping, except that one drive is reserved to store parity information. RAID 3 is not commonly used because its optimum performance depends on having synchronous data access to all disks (parallel I/O).


(Independent Access with Dedicated Parity Disk)

RAID 4 is similar to RAID 3, but emphasizes performance for a different type of application. It performs better for smaller file access large sequential file. However, RAID 3 and RAID 4 has an inherent bottleneck on the dedicated parity disk, it is also rarely implemented.


(Independent Access with Distributed Parity)

RAID 5 is similar to RAID 4, except that the parity is also striped across all the drives. For large data read/write functions the performance is same as RAID 4, but the parity disk is not a bottleneck, smaller I/O functions do not affect performance. Due to trade off between performance, fault tolerance and cost, RAID 5 is probably the most common RAID implementation


(Stripe Set of Mirrored Array)

RAID 10 is a combination of RAID 0 and 1. In this type of implementation a RAID 0 stripe set of the data is created across 2-disk array for performance benefit. A duplicate of first stripe set is then mirrored on another 2-disk array for fault tolerance. While this configuration provides all the performance benefit of RAID 0 and the redundancy of RAID 1, this level is very costly to implement because a minimum four disks are necessary to create a RAID 10 configuration.