5.1 Generic Requirements for SAN Management

only for RuBoard - do not distribute or recompile

After the SAN has been designed and built, the administrator will need to be able to manage all the equipment now associated with the SAN. There are three major areas of management that need to be considered :

controlling
monitoring
servicing

Control in the sense of SAN management refers to the ability to modify the state of the SAN resources. For example:

configuring a disk array
assigning LUN associations
setting levels of security

SAN management control must be able to provide heterogeneous node support. Storage products used in today s networking environments are likely to be from different manufacturers, but must work together on the SAN. One of the purposes of the SAN, after all, is to pool resources in order to reduce expenses and confusion.

The control portion of SAN management must also provide a single system image. The administrator must be able to view and modify all devices connected within the SAN from a single point. This would include devices such as:

storage devices

- disks

- tapes
servers

- UNIX

- NT
infrastructure components

- hubs

- switches

Having a single point of control also provides for better security of the management of the SAN. And finally, the addition and removal of nodes to the SAN should be accomplished without disruption to the network.

Monitoring is the ability to observe the state of the SAN. This would include the ability to understand the relationships between:

storage subsystems
server nodes
fabrics
allocations of assets within the network

As with control, monitoring requires the ability to see the entire SAN in order to inventory hardware and to view and manage capacity, use, and performance. Monitoring is a nondestructive activity, and therefore typically doesn t require as high a level of security as the control function. Monitoring should also provide a centralized capability. Centralized monitoring is essential for managing a large SAN extended over many kilometers or across campuses.

Service refers to the activities of:

finding and resolving problems
performing preventive maintenance

For example, servicing a SAN requires the ability to diagnose hardware problems on components within the SAN without interfering extensively with throughput. Also, adding and removing nodes should be accomplished without disrupting the SAN. This means that not only should the troubleshooting procedures be transparent to the user , but the design of the storage network should support the removal and replacement of components without interrupting data flow.

only for RuBoard - do not distribute or recompile