Physical Storage

Apply for Business Intelligence Certification Now!!

While implementing an application, the physical storage is always a main concern. Either the physical storage is available on server system with access rights or it will be provided locally and keep on replicated with server system. The physical storage must be managed from a single point of control. We are going to discuss two technologies, i.e. SAN and RAID.

SAN (Storage Area Network)

Storage area network designed to handle large data transfer over a local area network. Using high-end servers SAN supports data storage, replication and retrieval of data from multiple disk arrays. The SAN is dedicated network used to provide accumulated, block level data storage. Since SANs has its own network system of the storage devices which are not accessible by other device on the local area network (LAN). SAN system is flexible and can be integrate with the existing communication technologies like Fiber Channel or IBM optical fiber ESCON technology. SAN supports storage device monitoring, backup and restore of data, archival of data and retrieval of archived data and also supports data migration.

RAID

RAID expands as Redundant Array of Independent Disks. It falls under the category of disk device which is usually combined of two or more drives to improve the performance and for better fault tolerance. RAID technology is mostly implemented on server system and is not advisable for personal computers. RAID technology provides an efficient way to store redundant information without damaging overall network performance.

David Patterson, Randy Katz and Garth Gibson, in 1987 published their own research paper on “A Case for Redundant Arrays of Inexpensive Disks (RAID)”. This research paper is base of establishment of modern RAID technology. There was only 5 level of RAID defined in this research paper but the modern RAID technology has more then that.

Different RAID Levels

Level 0: Striped Disk Array. In this level a file is stored in multiple disk but not as replica. It is break into chunks and stored in multiple disk. Thus in simple words at this level the spreading of block of files into multiple disks without replication. This improves performance but is not able to deliver fault tolerance.

Level 1: Mirroring and Duplexing. In this level, monitoring of disk is performed. This level provides twice the read transaction rate of single disk and same the write transaction rate.

Level 2: Error-Correcting Coding. In this level the block level data is striped into bit-level data.

Level 3: Bit-Interleaved Parity. In this level each bit-level data is provided with dedicated Hamming-code parity.

Level 4: Dedicated Parity Drive. In this the data is striped to block level data (i.e. level 0) with a parity disk. If the data disk crash or failed, the parity data is used to recreate replacement disk. The disadvantage of this level is that it creates write congestion or blockage.

Level 5: Block Interleaved Distributed Parity. In this level the data and error correction information are striped to byte level. Thus there is one parity block per striped. This results in high performance and good fault tolerance.

Level 6: Independent Data Disks with Double Parity. It is similar to level 5, the only difference is there are two parity block per striped. The advantage of two parity block per stripe is that it can survive the loss of two disks.

http://www.vskills.in/certification/Certified-Business-Intelligence-Professional