I've always stressed on this forum that performance testing should not be an afterthought. Meaning, you should test the bare metal hardware using a live cd or a light installation of linux.
I generally use DRBL to boot an Ubuntu image I've loaded with all the drivers, management tools (e.g LSI/Areca/Adaptec) and storage testing tools (Iometer/dynamo, xdd).
Knowing what your raw LUN performance is a critical baseline metric.
That being said, Iozone is significantly different than dd. For one, Iozone can utilize multiple threads and reach a higher queue depth.
If your SAN is used for Virtualization, your concerns should be how well it performs random IO. Throughput tests are great for bragging rights, but don't typify what you'd see in real life applications. Most OS access patterns are mostly random of varying block sizes and like 60 - 80 % reads. Fire up a Windows VM and load IOMeter and use the file server or web server access pattern. This will give you an apple to compare to other apples. Sites like tomshardware and anandtech frequently test new hdds and raid controllers. You can see how you stack up.