Level 2: Switch-Level Redundancy In a two-switch cluster, as shown in Figure 23-2, each switch has the same set of storage devices and servers connected to it. This provides two independent data paths between each server and storage through separate switches. The second path is usually a standby and activates itself if the first path fails. The transfer of activity from a failed path to the standby is transparent to the user. Two SAN zones must be maintained within the fabric. Although the switches are connected to each other using ISLs, these ISLs should not be used under normal conditions for data transfer. Doing so increases the number of hops from one to two. The ISLs are used for the following tasks: ■■ They help simplify fabric management by using the same configuration copy for both switches. ■■ They are used for I/O transfer if two links fail — one from one server to a switch, and the other from the other switch to storage arrays. In that case, the ISLmust be used for I/O. The Xs in Figure 23-2 illustrate such path failures that necessitate ISLusage for data transfer. Server Acan still communicate with the storage devices by using ISLs to go from Switch 2 to Switch 1, and then to the storage devices.