We went from HP EVA P6300 to Nimble, were considering 3par for our upgrade as well. I have always liked HP's hardware but their software has always bugged me. We weren't even maxing out the EVA so perf wasn't highest consideration.
Drivers for me to choose Nimble over 3par were
- integrated out of the box perf monitoring; both on the array, and deeper analysis and proactive support via Infosight. we would have had to pay extra for that with EVA/3par
- small one: free vmware plugin, you have to pay extra one way or another for that with HP
- no tiering nonsense (ie if I want to dedupe then I gotta pin the workload to flash and then complex math figuring what size and how many luns to create at what tier for which application). With Nimble it all goes as fast as possible provided you get correct array and right amount of cache. I can still manage performance to a degree with Storage I/O control in vmware if I want.
- easy to understand all inclusive quote on Nimble's side (vs a 2 page set of dozens of line items for new 3par gear)
- cleaner simpler interface. the 3par gui didn't really make an impression on me and I am already biased against HP software
- customer referral calls: Nimble customers of our size were much happier about their product
- I really liked how the product has been built from the ground up to take advantage of spinning disk and SSDs strengths and mitigate their weakness' AKA CASL/NVRAM, and able to scale more easily in three directions. I got the feeling HP is bolting stuff onto a product designed for spinning disk yet with 3par. IE why are they using an expensive ASIC to do compression and dedupe when you can easily do that with commodity processors?
- lower TCO. with 5 year premium support I am still 20% below what it would have cost for 2 tier 3par for 3 years.
- decreased possibility of another forklift upgrade in my future
- easier to setup on my own, likely would have had to buy HP support to setup and optimize a 3par. nimble did actually send out somebody who helped at no cost once I had it in place to make sure I crossed all my Ts
- final straw was HP reseller FUD about how moving a sharepoint workload to Nimble at a former 3par customer's site killed performance for the entire array. when I followed that claim up with the customer it was completely bogus.