.SAN design reference guide Vol. 1-5 785350-001

A complete design specification includes the following elements:
• Topology map—Shows the logical SAN topology and fabric interconnect scheme; describes
a strategy to accommodate expansion and technological advances
• Configuration layout—Shows the physical layout of components; use for troubleshooting and
to verify the correct connectivity
• Storage map—Defines the storage system configuration and settings, such as host LUN
allocation and RAID levels
• Zoning map—Defines the communication access settings for devices and user ports in the
SAN
SAN topology
This section describes SAN features for enterprise-level SANs:
• “Multi-fabric SANs” (page 384)
• “Failover protection” (page 384)
• “Data access patterns” (page 385)
• “ISL ratio” (page 385)
• “Incremental SAN expansion” (page 385)
Multi-fabric SANs
HP recommends using two or more separate fabrics for enterprise-level SANs. Multiple fabrics
protect against potential failure points, such as hardware, software, or operator error. The failure
of one fabric does not affect other fabrics in the SAN.
SAN and fabric monitoring
For all single and multiple fabric configurations, HP recommends that you utilize intelligent fabric
monitoring tools such as HP Intelligent Infrastructure Analyzer Software (IIAS),“SAN infrastructure
monitoring” (page 404), B-series Fabric Watch and C-series RMON (see the user documentation
for your switch). These products provide detailed monitoring of individual Fibre Channel ports,
notifications, and in some cases, automated isolation of inoperative devices. This can help you to
identify abnormal conditions and avoid operational degradation. This degradation can adversely
affect operation, but does not necessarily result in a failover event.
NOTE: These tools are effective in detecting and avoiding most abnormal conditions. Failing or
marginal hardware can cause rare conditions which are not detected. In order to minimize these
rare conditions, HP recommends that you implement proper cable management practices and
end-to-end SAN monitoring.
Failover protection
Use failover technology in SAN configurations that have two or more fabrics. Each server has two
or more HBAs. If the communication path from one HBA to the storage system fails, the I/O traffic
is rerouted through the other HBA.
To minimize the risk of uneven workloads, configure the separate fabrics for similar size and
topology. You can also use failover protection in SANs with only one fabric to protect against
HBA, path, and storage controller failures. For more information, see “Data availability” (page 38).
384 Best practices