Please let us know how useful you find this tip by rating it below!
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
These days, network administrators have multiple choices when it comes to storage mechanisms. Terms like DAS, NAS and SAN tend to get thrown around a lot, but I run into lots of people who don't really understand how these different types of storage differ from each other, or when it is better to use one type of storage than another. In this article I will explain SANs.
A SAN is a network in which the storage mechanism (usually a disk array or tape library) is not directly attached to the server. Although direct attached storage (DAS) often offers better performance than a SAN, a SAN allows multiple servers to share a common storage mechanism.
A key limitation to a SAN, though, is that the storage mechanism is divided into virtual devices. Each server connected to the SAN owns one or more virtual devices. This means that the virtual device behaves much the way it would if it were physically attached to the server. In a SAN environment, you will never have a situation in which multiple servers are connected to a single virtual device and are using a common set of files.
Is it real, or is it virtual?
A virtual device is a sub-set of the storage medium. If the storage medium is a disk array, then a virtual device is usually a partition. If the storage mechanism is a tape library, then each individual tape drive might be treated as a virtual device.
The technical term for these virtual devices is a LUN (Logical Unit Number), which works the same way on a SAN that it would in an individual computer with a collection of SCSI devices. In fact, SANs access data from the storage mechanism in a similar manner to the way that data is accessed from a SCSI hard disk.
Networks that use Server Message Blocks (SMBs) or Network File System (NFS) perform network access at the file level. For example, in a Windows Server environment, the instructions for reading a file would look something like this: server_nameshare_namefile_name. However, to read the same file off of a SAN, the command would look something like this: Controller 4, Disk 2, block 6254.
As you can see, servers communicate with a SAN device at a low level. Rather than making calls directly to a specific file, calls are made to a SAN at the SCSI level, just as if the disk were physically attached to the system.
Suppose that a particular SAN consisted of a RAID array filled with individual partitions. You might wonder what the advantage is of using a SAN as opposed to DAS. There are actually several advantages. One is cost. If you have a SAN consisting of several servers and a single disk array, you have saved some money because you did not have to purchase a separate disk array for each server.
Another advantage is flexibility. In a SAN, the servers are logically attached to individual LUNs. This means that you can re-assign a LUN to a different server without having to rewire anything. LUNs can be assigned on an as-needed basis.
Fiber channel vs. Internet connection
Besides servers and the storage mechanism, there is one more component to a SAN. You need a way of physically connecting a server to the storage mechanism. At the moment, the most common type of SAN connectivity is fiber channel.
A fiber channel SAN is made up of a number of fiber channel switches, all linked together to form a network. This type of network is known as a fabric. In a fiber channel SAN, SCSI commands are sent from the server, over the fiber channel network, to the storage mechanism.
A more recent SAN technology is Internet SCSI, or iSCSI, a protocol that allows SCSI commands to be sent through the TCP/IP protocol. The advantage of iSCSI networks is that they can function over the Internet or over an Ethernet network, rather than requiring a fabric network.
SANs are primarily used in large enterprise environments. At the moment, SANs are expensive to create, but iSCSI is starting to bring the price down. Network-Attached Storage (NAS) tends to cost less than a SAN, but most SANs perform better and are considered more reliable than a NAS device.
Brien M. Posey, MCSE, is a Microsoft Most Valuable Professional for his work with Windows 2000 Server and IIS. Brien has served as the CIO for a nationwide chain of hospitals and was once in charge of IT security for Fort Knox. As a freelance technical writer he has written for Microsoft, CNET, ZDNet, TechTarget, MSD2D, Relevant Technologies and other technology companies. You can visit Brien's personal Web site at http://www.brienposey.com.