In many situations, it is difficult recommend SAN technology. The advantages of central management, ease of backup, better disk utilization are clear, but availability is not improved by SAN vs DAS. And the reason for this is that SANs do fail, and when they do, they cause an entire data center to fail. When a SAN goes down, the storage and UNIX teams are heavily overwhelmed with incidents to manage and resolve. All of the applications are down and all the application support people are at the mercy of the storage and UNIX teams to get all the systems back up. It is a single point of failure for an entire network. It’s better to construct the management and redundancy as part of the application with the recognition that media will fail. Constructing robust fail-over and redundancy in the application and avoiding SAN infrastructure will make disk failures trivial matters that don’t impact more than one machine. The LDAP and SMTP protocols and the servers that server LDAP and SMTP are very redundant at the protocol and application layer. Designing LDAP and SMTP infrastructure where no machine is a single point of failure is not terribly difficult, so flash memory and even ramdisk memory should be game.

© Copyright 2020 Rex Consulting, Inc. – All rights reserved