Storage Area Network: Architecture

LUN Masking

The purpose of a SAN is to allow multiple servers to connect to a single shared storage device. A SAN allows you to group the hard drives in the shared storage into one large or multiple smaller logical units and assigns a logical unit number (LUN) to each logical grouping.

Each LUN is then assigned to a given host connection where appropriate. This allows large storage devices to be broken down into manageable pieces. LUNs can also be used in part of the failover process when a component in the storage path fails.

LUN masking is the ability to allow specific LUNs to be exclusively assigned to and accessed by a specific list of hosts. It is possible, with LUN masking, to reliably attach a single LUN to a single host connection. Other hosts will not be able to access the LUNs to which they are not assigned. This is critical as during system boot up, each server in the SAN attempts to use all storage that is physically connected. In this case, LUN masking prevents the servers from accessing the same data at the same time, which could cause file corruption.

There are several places where LUN masking can be implemented, two of which will be explored.

Server Based LUN Masking

LUN masking on the server level requires a software package to be installed on each server. In this example, the host system will still see all of the LUNs, but will essentially ignore those LUNs not assigned to it.

The drawback to this type of LUN masking is that the software on each server will have to be maintained.

Storage Based LUN Masking

Storage based LUN masking is a feature of the storage system itself. There is no need to install software on each server, and a server will see only the LUNs that are assigned to it, instead of all the LUNs as in server based LUN masking.

The fact that software does not have to be maintained on each server is an added benefit to this type of LUN masking, since there are generally many more servers in a SAN than there are storage devices.