Clustered file system

Poll without page-refresh

	del.icio.us
	Digg
	Furl
	Reddit
	Rojo
	Add to OnlyWire

A clustered file system is a file system which is simultaneously mounted on multiple servers. There are several approaches to clustering, most of which do not employ a clustered file system. While many computer clusters don't use clustered file systems, unless servers are underpinned by a clustered file system the complexity of the underlying storage environment increases as servers are added, effectively limiting these clustering styles to 3 or 4 servers for all practical purposes.^{[citation needed]}

1 Shared-disk cluster file systems
2 Shared-nothing clustered file systems
3 Usage
4 See also
5 References

Shared-disk cluster file systems

The most common type of clustered file system is the shared disk file system, in which two or more servers are connected to a single shared storage subsystem, such as a stand-alone RAID array or SAN. Examples of this type of file system would be VMFS and GFS.

Shared-nothing clustered file systems

Another clustered file system approach is to have each node use its own local storage, and communicate data changes to other nodes via some network or bus. In this case disks are not shared amongst nodes, but are instead dedicated to a single node and made readable and writable to other servers participating in the cluster. A single global file system is presented as a result of the cooperating nodes. Examples of this type of file system would be the file systems in Isilon and IBRIX clustered NAS products.

Usage

Scientists working at ALICE will be using a 4 Gbit/s Fibre Channel SAN with a clustered file system to store the massive amount of data generated by the experiment (estimated at 1 GB/second for a month). Reasons quoted for this choice include "performance, scalability and vendor independence"^[1]