2 posts • joined 14 Jul 2011
Wow, haters galore - incredible...
Go beyond blogs and wikipedia for information before you speak - read an analyst report every now and then. Rampant ignorance here….
@thegreatsatan - all of it. It's being finalized and released as we speak - once it's GA, then the field is trained. This is common practice. You talk like every other all flash array has VAAI when none of them do. We're the first and have the most comprehensive support. Agree about the Kaminario comparison, they're a shared DRAM and caching appliance with some flash mixed in. We're a standards-based, accessible and serviceable all flash array.
@chris coreline - 650k usable IOPS for the common man, 6 GB/s usable throughput, plugs into anything, stupid easy to manage, centrally manage 72TB of all flash capacity, zero single points of failure - this is all that matters, this is what the industry is missing. No one else has all of this (with a customer base currently in the field). Customers don't need more storage "cupholder" features when it comes to storage performance. Performance, stability, and availability first, bells and whistles later...
@chunky lafunga - we've been shipping these units for four years with AutoSupport enabled. VDI is up to 99% write and we have VDI customers pushing multiple sustained TBs per DAY through even the smaller units over the course of YEARS. In fact, on the INVICTA user interface, you'll be able to monitor drive wear in real time. Whether you believe it or not, it's real - these drives last longer than 7 years even under a full daily write workload. Visit our side for a video demo of the UI.
@anonymous coward - the 1.5TB unit delivers 250,000 IOPS 100% random, 100% write, sustained at 4k over 10 GbE iSCSI. For $50k. And that's at RAID 5 with a hot spare. First of all, you can't even get that kind of write performance from anywhere else, and if you could, not for less than millions of dollars (or pounds, or euro...). Second of all, visit our website for the ESG third party performance lab validation report. No one else in the all flash array space has third party lab validation. Third of all, hard drives can get you about 200 IOPS each (less after RAID), so you're looking at about 2k IOPS per $30k tray. It's not even a contest.
@morg: see above for 7 years being a crap number. Actually, it's more like 20 years. As for performance being last gen, compared to who, or what? I'm not sure if anyone every taught you this, but YOU CAN'T BELIEVE EVERYTHING YOU READ! Check the SPC-1 performance benchmark record - the current storage performance title holder is a 10 rack $4 million, 2,000 hard drive monster from IBM and it does about 520k IOPS. And it's not even GA. Go to our website and click on the video showing 650,000 IOPS in 14U drawing less than 2,000 watts of power. Last gen? What's now gen? What's next gen? Silly.
My email is firstname.lastname@example.org. Please feel free to send me an email with any questions before posting some bitter, ignorant rant.
Context of Hot Swappable
The comment Dan made about hot-swappable, properly put into context, was more about the nature of WhipTail than it was about Violin.
WhipTail is architected in a familiar manner - drives in the front, plugs in the back. On the front of the array it is simple to unlock the front bezel and see the 24 drive slots that can be populated with anywhere from 1.5TB to 12TB of hot-swappable MLC based drives that can be replaced, upgraded online, etc.. Access is easy and natural.
Violin on the other hand manufactures their own memory modules implemented within a locked-down glued-shut appliance. So while it may be "technically" possible to hot-swap a memory module, it certainly isn't hot-swappable in the general industry understanding of the term.
That's like saying a human kidney is hot-swappable. Technically true, you don't have to stop the heart and resurrect the patient to get the job done, but it's not as hot-swappable as, say, a pair of shoes.