You could tell that YES, there was a bit of “proud parent” dynamic…
… And YES, we have a ton of install base data (on data reduction rates, on durability, availability and more here)…
… And, yes, it is a monster in a sense – packing a lot of punch in an amazing package in the 4.0 software update and the new more dense package.
… And yes, customers are indisputably voting with their feet and their dollars – IDC published their “Worldwide All-Flash Array 2014-2018 and 1H14 Vendor Shares” doc. We’ve gotten the rights to republish sections – click on the link to the right. If you want the full doc, it’s here, but you’ll need to get it from IDC.
… But then I looked at the customers below, and realized, I totally missed the main point – watch it, then read on…
The point they reminded me of is that XtremIO is a “Beauty”, AND a “Beast”.
New XtremIO 4.0 goodies
|Performance economics||Simple. Fast. Always. Period. “The easy button”.||BRUTALLY FAST – consistent sub 1ms latencies, loaded to the gills, all features working (including the new remote replication) – and with the new scale-out limits increased – an easy 1M IOps – not in a fake workload designed to maximize results, but to reflect as close to real world, loaded with data, and with all features in use.|
|Capacity economics||Dense 40TB X-Bricks coupled with always inline data reduction you can bank on = great $/GB for broad sets of workloads – anything transactional.||1.6TB eMLC SSDs coupled with the new monster 40TB X-Bricks = we’re continuing to ride the NAND curve, and yes, if you look at the durability data I published here which show that no customer has less than 20 years left of wear – you bet we can use even denser, lower cost NAND thanks to industry leading write amplification avoidance.|
|Features||There is nothing simpler than XtremIO, and the features are just “there” and work. It’s easy.||Crazy-rich snapshots, remote replicas, consistency groups, RPOs in the sub 60 second ranges, able to cope with huge change rates. Improved resiliency. Data Reduction that is inline, all the time – because of in-memory mapped metadata – in a scale-out design.|
|Scaling||Starts SMALL. 3TB. Scales out so scales as you need.||8 X-Bricks with 40TB each is a 1PB (with data reduction) in a single “array”, and single floor tile. Scales like a mofo. And it’s real scale-out: every IO can be served by every X-Brick, with a fully utilized active-active end-to-end system design.|
|Inline, all the time.||Shouldn’t features work like this? Always on, always inline – they just “are” in the same way that the sky is blue.||All data services are inline, all-the time. Thin provisioning, deduplication, compression, data-at-rest encryption – all happening on a per I/O basis without fail.|
To some – they admire “beastliness”. Others admire “beauty”. There is of course – the rare times where something is BOTH beauty and the beast.
Last week I was in St. Louis – visiting a ton of great customers, and also speaking at a EMC/VMware sponsored Women’s Leadership Forum (incredible panel!).
I ALSO visited World Wide Technologies (WWT) – an amazing EMC (and VMware, and Cisco, and a huge long list) partner.
They have an Advanced Technology Center which is amazing. They can support real PoCs with customer’s real workloads – and due to their partner ecosystem – it’s not just one vendor – they can literally assemble almost ANY PoC combo.
Here’s a picture of XtremIO in their ATC, and while 4.0 doesn’t GA until the end of June – they already have the bits (and have been helping us through the beta).
I framed the picture this way because beside it there are XtremIO’s competitors (our partners work in an open ecosystem) – and I don’t want to get our great partner in any trouble.
But – if you want to put us (and others) through their paces – don’t listen to me. Give it a shot. Reach out to WWT and You’ll see what I’m talking about – XtremIO 4.0 is a beast indeed – but simple, starts small, and is beautiful.
As always – comments welcome!