Specifications
DATA CENTER BEST PRACTICES
SAN Design and Best Practices 11 of 84
Servers and storage devices should be connected to both networks utilizing some form of Multi-Path I/O
(MPIO) solution, such that data can ow across both networks seamlessly in either an active/active or active/
passive mode. MPIO ensures that if one path fails, an alternative is readily available. Ideally, the networks
would be identical, but at a minimum they should be based on the same switch architecture. In some cases,
these networks are in the same location. However, in order to provide for Disaster Recovery (DR), two separate
locations are often used, either for each complete network or for sections of each network. Regardless of the
physical geography, there are two separate networks for complete redundancy.
In summary, recommendations for the SAN design are to ensure application availability and resiliency via
the following:
•Redundancy built into fabrics to avoid a single point of failure
•Servers connected to storage via redundant fabrics
•MPIO-based failover from server to storage
•Redundant fabrics based on similar architectures
•Separate storage and server tiers for independent expansion
•At a minimum core switches should be of equal or higher performance compared to the edges.
•Dene the highest performance switch in the fabric to be the principal switch.
Switch Interconnections
As mentioned previously, there should be at least two of every element in the SAN to provide redundancy
and improve resiliency. The number of available ports and device locality (server/storage tiered design)
determines the number of ISLs needed to meet performance requirements. This means that there should be
a minimum of two trunks, with at least two ISLs per trunk. Each source switch should be connected to at
least two other switches, and so on. In Figure 2, each of the connection lines represents at least two physical
cable connections.
fig02_SAN_Design
Redundant
HBAs
Dual-attached
hosts
Dual-attached
storage
SAN A
Fabric
SAN B
Fabric
Redundant
controllers
Figure 2. Connecting devices through redundant fabrics.