Storage Area Network

Definition

A SAN is different from the other storage systems such as the NAS (Network attached storage) by a low access level to the discs. To simplify, the traffic on a SAN is very similar to the principles used for the use of the discs Interns (ATA, SCSI). It is a mutualisation of the resources of storage.

In the case of NAS, the resource of storage is directly connected to the Ethernet network of the company. Waiter NAS integrates the support of multiple filesystems network, such as CIFS (Common Internet Spins System), the protocol of file-sharing of Microsoft, NFS (Network File System), a protocol of Unix file-sharing or AFP (AppleShare File Protocol), the protocol of file-sharing of Apple. Once connected to the network, it can play the part of several divided file servers.

In the case of the SAN, the bays of storage do not seem the volumes shared on the network. They are directly accessible in block mode by the filesystem from the waiters. In light, each waiter sees the disk space of a bay SAN to which it has access like its own hard drive. The administrator must thus define very precisely the LUN (units logical) and the zoning , so that an Unix waiter does not reach the same resources as a Windows waiter using a different Filesystem.

Advantages

One of the first interests of the SAN is more not to have to be concerned with make evolve/move the quantity of discs formerly reserved for a particular waiter or to be afflicted space not used of an oversize waiter in disk space.

The disk space becomes almost unlimited and is evolutionary at will by the addition of discs or of bays of storage on ENV. the storage space mutualized for the waiters allows to optimize the management of the discs and to make easier the data back-ups.

Another advantage of the SAN is the possibility of implementing functions of replication (copy of data synchronous or asynchronous between two bays) and of snapshot (duplication of a volume to use it on another waiter or to safeguard it for example). These functions make it possible to make safe the data (physical establishment in distant buildings) and to optimize the availability of the applications.

Types of SAN

The SAN are built with an aim of providing fast and reliable disk space. The most widespread technology to reach that point is the use of the protocol Fiber Chanel which makes it possible to reach high flows.

The fiberoptic equipment being very expensive (switchs in particular), two new technologies made their appearance:

ISCSI : they are the same orders SCSI but on TCP/IP.
AoE : here it is ATA on Ethernet. This technology does not make it possible to make Routage.

Compatibility

The main difficulty in the design and the exploitation of SAN is to make sure constantly of the compatibility of the various components which are implemented: charts Fiber Chanel with their version of Firmware, pilot (driver), possible software of multipath, firmware of the switchs, version of controller of bays of storage. It is fundamental to respect the recommendations of the manufacturers or the integrator (matrix of compatibility) to guarantee the good performance and the performances of the applications lodged in the ENV. one.

SAN in action

Still little time ago one found SAN only in the great data-processing centers of companies needing a great volume of very protected storage. The prices becoming more reasonable, with the more modest data-processing centers are equipped with simpler architectures SAN (as represented in the diagram below) .

A larger site will set up a more complex architecture, made up of several SAN connected to each other. Each SAN which can be equipped with a great number of bays of discs which can be heterogeneous

See too

Fiber Chanel: protocol more used for the SAN
ISCSI: protocol based on TCP/IP
Network attached storage: a more widespread alternative of storage in company.
RAID: methods of optimizations and securities of storage on discs
SCSI: the standard used to communicate with the discs

External bond

'''SAN Comparisons'''

Random links: