I’m super excited to announce that we have just GA’d XtremIO X2 during VMworld 2017 and that’s no coincidence!
XtremIO is the best AFA for your virtual workloads and it’s time to sum it all up in one post so here goes
The AFA vendors promised us that by moving to their box, we could consolidate their entire virtual workloads into one or two frames but that is hardly the reality, most of the dual controllers based AFAs out there cannot cope with the load of some of the heavy virtualized apps that are out there so you actually end up with more silos
And if the fact that your dual controllers AFA architecture cannot cope with the IOPs/latency, you also have another issue, they don’t enough RAM to be used as the metadata for the actual data, think of metadata as the DNA structure of your actual data, the faster it is, the faster you know how to retrieve it, that was always the magic behind the core architecture of XtremIO
Enter XtremIO X2.
if you haven’t read the series of blost post we wrote on X2, starts here
X2 design goals were
Performance – to provide you NVMe like latency with the cost of traditional SSD’s
We were able to achieve this by analyzing our call home data gathered from thousands of clusters out there and finding out most of the IOs are coming from small IOPS
We could have just used whatever is the latest intel CPUs available at the time but we wanted to achieve much higher numbers than what was achievable with them and so we came with a new feature called “write boost” which improve both reads and writes latency by up to 4 times!! You can find a much more in-depth explanation in a blog post + demo I wrote here
you can also watch a video recording i had with our chief architect here
want to see the effect of write boost + scale out in a virtual environment? here’s a demo that will explains it
- improving upon all the AXIs of the X1 success
Cost. in X2 we now support / have:
Larger drives (1.92TB and 3.68TB).
Denser DAE (from 23 drives to 72!, start with 18 drives and grow with packs of 6).
25% better compression (in some cases, its more than this).
Both odd / even configurations are supported.
Write boost which improve the $/IOPs.
No more BBUs which improve the usable capacity per a floor space.
The new UI is HTML-5 based but that’s not it’s highlight, it’s the intelligence we put in the UI which is unique in the industry
things like weekly pattern (think VMware vCOPS)
And block size per IOPs and histograms reporting which are the holy grail where it comes to storage reporting
See a demo here
- enterprise grade features
Native replication which is going to change how you replicate data, think Avamar Vs data domain where all the changes are calculated at the source, see a post and a demo here
or just the demo here
Best VMware integration
The VMware integration we have is 2nd to none and I’d like to think of it as pillars of different hookup to VMware vSphere
The 1st one is the “CORE”, this include a unique VAAI XCOPY integration which clones your VMs in matter of seconds instead of minutes, all the data services of XtremIO are inline
As of vSphere 6 U3, there is a native SATP adapter that detects the fact you are using XtremIO and will change the default SATP from “fixed” to “Round Robin”
The 2nd one is “Manage & Monitor”, this one includes the VSI vCenter plugin and the ESA (vROPS) plugin, both are free and both are giving you the options to completely manage your XtremIO array from vCenter and to monitor it using VMware vRealize Operations, see demos here
the 3rd one is “Protect”, it’s allowing you to fully protect your Datastores, VMs, files from within the VMs and to apply the Copy Data Management (CDM) on them, see a demo of Appsync here
the 4th one is “automate”, this is where our customers are going, they want to automate everything and as such they can use our free vRealize Orchestrator XtremIO plugin, see a demo of it here