ABC.com plans to roll out several experimental large-scale host clusters with low-cost blades for
computationally intensive workloads. All IS operations are in a single data center (50 m x 40 m).
They use direct-attached storage on all hosts. For better manageability of storage, they are
considering a SAN infrastructure with CLARiiON storage (CX3-40s). Preliminary design calls for
200 server blades. Cost and performance are both critical considerations, but high availability is
not. Your initial design has single-pathed FC connectivity to each host. Under sustained load, each
server blade can sustain up to 4,000 IOPS at 16 KB per I/O. Which additional information do you
need before you can attempt a preliminary, high-level SAN design for this environment?

A.
Server model/make, OS type and revision, clustering software

B.
Breakdown of I/O traffic by customer application

C.
Customer’s current disaster recovery strategy

D.
Expected cache hit rates on the storage arrays