Well said. When you stuff 70TB of raw flash in 3U, de-dupe doesn't have to exactly be in your playbook, especially in the database workload market where nobody is thinking they have to have de-dupe. However, all of the "me too" players who are building smaller, cheaper boxes with less flash totally polluted the reality of the value of de-dupe to the customer, so Violin needs this feature set to speak to customers competitively. Right or wrong, customers have it in their head. They have 20 vendors saying de-dupe matters, and only Violin was saying it is not critical. Both sides were partially wrong, but for purposes of their own agenda.
Ironically, all the "me too's" had to use de-dupe to compete with Violin initially because their boxes have way less flash inside. Look at their spec sheets. Nobody else is selling anywhere near 70TB in 3U. De-dupe gave the "me-too's" an opportunity to maximize the effective capacity of their smaller box, and then it became a thorn in the side for Violin. That game is over now, so the next 4 quarters should be interesting for Violin in de-dupe centric work load deals.
If you are a massive company with performance on your mind, you just buy tens of arrays from Violin and go back to sleep. You don't even think about de-dupe, but so many data center decision makers are so misinformed on the all-flash thing. Violin's biggest problem is their arrays are lasting so long in the data center. They need to play that 3-year upgrade cycle game like everyone else instead of selling awesome technology that will last for 7-10 years in the data center. If they came up with a sh!tty and cheaper product, they would be doing much better. Look at Pure kicking ass with a sh!t product.