Observations from the Flash Memory Summit 2014, Part 2, beyond the feature checklist

At the Flash Memory Summit, Josh Goldstein (our VP marketing and products management) was on a panel discussion with some of the other AFA vendors.  Attached is what he presented and you can download the other vendors from the FMS webpage below


Part 1 of the series which you can read here: http://itzikr.wordpress.com/2014/08/18/observations-from-the-flash-memory-summit-2014-part-1/  was a reflection about being chosen as THE market leader.

this post is a reflection about the storage panel that was held during the summit..

let’s start by saying that it’s interesting how the lack of a proper (some may call it “true”) AFA specific architecture is being compensated by some clever marketing pitch


  • Overemphasis on 100% performance all the time.  some vendors will try to make a big deal of losing performance when a controller fails.  while they are active/passive – a VERY simple implementation that does maintain the same performance with a failed controller, but which doesn’t deliver close to XtremIO’s performance the other 99.999% of the time.  Which bring us to….


  • ….1M IOPS – some competitors will try to say this doesn’t matter and that you should focus on consistency.  Well, we wholeheartedly agree you should focus on consistency and predictability of performance.  And that you also need 1M IOPS for where applications are heading and for workload consolidation.  Just attend my upcoming session at VMworld to hear about real world customers who put in a PO in order to  give every developer a high performance database sandbox.  Can’t do that on a limited Active / Passive architecture that is limited in intself and that is before we are even talking about scale – out!
  • “Adaptive I/O” – because some competitors can’t compete with always-on, always-inline, consistent and predictable performance and data services.  So they’ve changed their messaging and now say their array is “adaptive”.  You know what that means?  It means the more load it contends with the less CPU time is allocated to process data services like deduplication and compression.  That leads to inconsistent and unpredictable performance.  Don’t let the nice marketing to sway you the customer away by proving out their array under low load situations.  We encourage people to hammer XtremIO and this is the reason some competitors hates when customers actually put the pedal to the metal with their array.  Insist that they do.
  • Defocusing scale-out.  some competitors will try to tell customers that nobody needs scale-out because customers prefer fault domains in cloud environments with multi-array management instead of scale-out.  We know this isn’t true.  Customers are pushing us for larger and larger cluster sizes TODAY! .Yes, there is a point about fault domains – but the domain size is much larger than what a single active competitive  controller can handle.  You still need scale-out, even with reasonably small fault domains.  Remember – other Storage arrays designed to run today’s workloads with some more IOPS.  XtremIO is designed to change the entire way you deploy applications and give you the power to run tomorrow’s workloads.


  • A competitor calls 4KB benchmarks “unicorn IO” and wants to focus the attention on I/O sizes of 32KB or larger – what they call the “real world application IO zone”.  BRING IT!  We will win over them on any benchmark at any IO size – especially on cost-equivalent systems.  We’ve proven this in POC after POC.  A fully populated $1.5M competitive product is no match for two X-Bricks.  And we still have the headroom to add 8 more active controllers!

Back to the Future Part II

  • some competitors  likes to say “not all data reduction is built the same” – WE AGREE.  Customers should demand inline all the time.  There’s no other way to get consistent and predictable performance and superior flash endurance.

DeLorean DMC-12

  • competitors are now trying to steal our “Architecture Matters” tagline.  Remind youself whose message that is and why we say it all the time, as well as the numerous architectural disadvantages other products have – active/passive dual controller, post-processing data reduction, cMLC and system level garbage collection, metadata only partially cached in RAM, etc.

Curbside collection

  • most of the other competitors  will try to say that there are many garbage collection strategies and that system level garbage collection isn’t bad – that customers should focus on results, not implementation.  WE AGREE.  We’ve proven in POC after POC that other vendors system level garbage collection wreaks havoc on their performance and consistency.






In summary:

Flash array designs vary widely.

Things that sounds the same aren’t.

If your’e only thinking of flash arrays for performance, you are missing much of their value.

Edcuate yourself and make informed decisions.

Don’t be fooled by hype or hero numbers.

Look for true innovation.

Test, test, test

Leave a comment

Filed under Uncategorized

Observations from the Flash Memory Summit 2014, Part 1




the Flash Memory summit took place this year between august the 5th till the 7th.

all proceedings from the Flash Memory Summit have been posted online and you can download them by clicking the link below



here’s some observations that are relevant to you as a potential buyer of an All Flash Array:

http://www.flashmemorysummit.com/English/Collaterals/Proceedings/2014/20140805_103C_Berry.pdf   – this presentation is from IT Brand Pulse, which surveys IT end users, and who for the 2nd year in the row named EMC/XtremIO as the most favored block all-flash array.  This is already covered on http://www.xtremio.com in the ‘Awards’ section, here are some helpful slides:

It shows EMC as the HOTTEST IT Storage Company.


  • Slides 24 and 31 show the categories in which EMC was voted at the top by end-users:


  • Who is #1 in this market?  EMC is!


  • And who do end-users plan to buy from?  That’s right, EMC!  Look at where the other competitors are.


  • Slide 54 –an interesting one, it demonstrate that you, the potential buyer, “evolved” quite a bit yourself but what does it mean? it means, that there was a lot of progress in the thinking process from buying a “dumb” AFA that provide performance only, to a more sophisticated array with data services, radically different architecture etc’, that is a very big thing as the AFA market  is still confusing to some in terms of the differences between different vendors products, different testing methods etc’, i will talk more about these points in the 2nd part of this series.



Leave a comment

Filed under Uncategorized

Mitrend For EMC XtremIO



EMC SE’s and Partners have been using Mitrend for years now, in it’s essence, it’s an Pre-Assessment / analysis tool that allow EMC and non EMC customers to look at their existing dataset and analyze what array is needed if migrating from say netapp to a VNX or a VMAX, the ease of use is the key metric here, no one wants to use a tool that then require you the partner to do the anlysis yourself, you want to know what array you need, how many drives you need and how many engines. SIMPLE.

Today, we are pleases to announce the EMC XtremIO support for Mitrend for capacity assessment, the initial release will focus on capacity and future versions (they are rolling fast!) will focus on performance analysis.

how do you use the tool?



1. in step one you log in to the portal, if you’re an EMC SE or a partner you shouldn’t have an issue, if you are a partner and still do not have an account, just speak to your EMC partner alliance manager and he / she will get you an account.



2. when you log in to the portal, you are faced with a simple workflow, the collection, submission and viewing the assessment



3. since this tool support many sort of assessments, you want to view the instructions for the XtremIO one


4. once you are there, you can email yourself the assessment tool or download it directly.

5. the tool can either run on a windows host (the exe file) or on a linux or vSphere host (the .bin file)





6. running the tool in a vSphere enviornment, you want to install the bin file on an esx that is mapped to the volumes (VMFS and / or RDM’s), and in order for you to tell the tool which volumes you want to scan, you first need to run the following command : “ esxcli storage vmfs extent list “

once you get the output of all the volume this ESXi host sees, you can then run the tool on all or the selected devices you want it to scan:

./DataAnalyzer –X /dev/disk/<device_name>;







7. the screenshot above shows you how the tool looks when it runs on a linux host (very similar to an ESXi host)


8. now it’s tike to create the assessment so we can generate reports based on it.




9. the last step is to upload the files, there’s pretty much every way there, simple upload, FTP, syncplicty (my favorite!) or even via dropbox.


10. that’s it!, after a while you will be emailed back with the report, the screenshot above and below show you some slides from the final report, awesome!


these reports are not relevant for vSphere, they are more relevant for file servers or even VDI VM’s!


below you can see the most important slides, the data reduction ones, already supporting deduplication and compression which are now part of EMC XtremIO







what’s coming next?

throttling and performance reports for XtremIO so you can actually know how many X-Bricks you will need not just from a capacity point of view but also performance.

so, what do you say, isn’t it a cool tool or not Smile


Filed under Uncategorized

The German EMC Forum 2014: It’s Time to Redefine IT


Do you want to be part of something big? Are you looking to get the knowledge YOU need to rise to the next platform?

Do you want more XtremIO? You are located in Germany? Attend the German #EMCForum in Mainz! Sept16. http://emc.im/Forum2014DE

The worldwide EMC Forum is allowing you to do just that. We’re giving you the chance to redefine IT, engage with our top technical experts and influence the community at one of our top free events this year.

You’ll have already begun to see how megatrends like Social, Mobile and Big Data are redefining the role of IT. They’re creating new business opportunities and threats, and ushering in a whole new IT era – the “3rd Platform.”

At EMC Forum 2014, we want to hear your voice, listen to your journey, and redefine IT and business together with you – helping you to get ahead of the competition. You’ll also get to see the latest solutions, talk directly with EMC experts and partners, and network with people just like you at the event.

So, why not help your organization:

  • Bridge the present and the future: letting new and existing applications run side by side
  • Redefine IT: embracing private and hybrid clouds, and creating a low cost, highly flexible software-defined data center
  • Redefine business: harnessing the full power of mobile and Big Data applications, to boost productivity and enable smarter, real-time decisions

Are you ready to get practical advice on how to thrive in the new era of IT? We hope so!

Find out more http://www.emc.com/forum2014

Be part of EMC Forum 2014

- Register now for your complimentary ticket!

- Follow the EMC Forum community (http://emc.im/EMCForum_ECN) and have your say: vote on ideas, read the blog, join the conversation and more.


Leave a comment

Filed under Uncategorized

Got OpenStack And XtremIO, You just got Cindered!


OpenStack is leading the open source cloud revolution and has become the de-facto open platform for managing private and public clouds

OpenStack allows storage resources to be located anywhere in the cloud and available for use upon demand

OpenStack reduces the complexity in managing and consuming resources across heterogeneous environments

With the availability of the XtremIO integration to OpenStack, customers can now easily connect their OpenStack cloud to an XtremIO storage array.

OpenStack architecture in a nutshell

•Fully accessed by common set of APIs and common Dashboard


•Supports heterogeneous resources

•Compute resources, Hypervisors, storage backends, networking resources

What is Cinder?

Cinder is the block storage service for OpenStack.

Cinder directs the creation and deletion of volumes on a storage array.

Cinder attaches/detaches volumes from instances/virtual machines (VMs) created by OpenStack.

Storage resource allocation is performed on demand based on the OpenStack cloud requirements

XtremIO Cinder Driver


XtremIO Cinder driver sends storage provisioning commands to the XtremIO array.

XtremIO Cinder driver implements the XtremIO REST APIs

The OpenStack cloud can access XtremIO using either iSCSI or Fibre Channel protocols.

XtremIO Cinder Operations support


Cinder Driver Setup & Installation


–Make sure XtremIO array is connected to the OpenStack compute resources (initiators)

–For iSCSI

▪Configure iSCSI portals in XMS

▪CHAP initiator authentication is supported. If CHAP initiator authentication is required, set CHAP Authentication mode to initiator in XMS.

—The CHAP initiator authentication credentials (usernames and passwords) are generated automatically by the cinder driver

–Set up a dedicated account for Openstack with Admin role privileges (recommendation)


–The driver should be installed on the cinder host that has the cinder-volume component

▪Copy xtremio.py to cinder/volume/drivers/emc

Cinder Configuration

Edit cinder.conf /etc/cinder/cinder.conf



volume_driver=cinder.volume.drivers.emc.xtremio. XtremIOFibreChannelDriver

san_ip=[xms IP address] à retrieve from show-xms CLI Command

san_user=[xms username] à the XMS user-name previously defined

san_password=[xms password]

Restart cinder



Volume Created in XtremIO


Create Snapshot in OpenStack


Snapshot created in XtremIO


Attach Volume to VM in OpenStack


Volume mapped in XtremIO




OpenStack Demo

wrapping it all up, you can see a video Iris Kaminer, a PM in our team made.

Leave a comment

Filed under Uncategorized

VSI (Virtual Storage Integrator) 6.2 is Out, XtremIO is IN!


On the heels on the Mega Launch event in which we announced XIOS 3.0, another product release has just been GA, the VSI plugin.

if you are an EMC customer and never heard about it, it’s really time to wake up, this (FREE!) plugin is really a framework for the EMC products for VMware vCenter, at it’s core, it allows you to view storage array metrics and provision VMFS datastores or RDM’s devices, present them to the ESXi hosts and all of this without using the storage GUI. this release is the first release that support XtremIO AND the vCenter web interface, prior releases supported XtremIO but only worked with the classic vCenter.

attached below are two videos in case you find the documentation too complex:

01. Installing and Configuring the EMC VSI 6.2


02. Provisioning Datastores And RDM’s



but this release has also another feature which ‘im proud to say was co-invented by Tamir Segal who is a PM in our XtremIO team and your’s trully.

it really started as me scratching my head around the new XtremIO Snapshots capability and how revolutionary it is (you can read more about it here:


my question was, why are we even using VDI full clones, the answer is of course, “because customers wants to have persistent desktops” but then my head start spinning, I was thinking, “well, we can still give you a persistent desktop that will look like a full clone but why waste XCOPY time or even logical capacity??” without the penalty of the VMware linked cloned technology (can’t replicate it, you always need to worry about the parent –> child relationship etc’ etc’) and so we started to toy with the following logic:

1. create one datastore that will have X amount of VDI VM’s in it (full clones, all cloned from the same replica)

2. snapshot this volume X amount of times.

3. present the snapshot to the ESX’s cluster and renaming the VM’s, run sysprep on them and upload their computer account configuration to the VDI broker.

4. repeat it on the other snapshots..

and so, we went with this request to the VSI team and they made it possible, so you see, I’m Xtremly happy now because this platform called “XtremIO” is so unique and it allow us to really think outside of the box and leverage technologies in a way that was never done before!

by the way, there is an even extra value in this plugin if you work with Citrix XenDesktop and EMC XtremIO since there is no native way in Citrix Studio to create full persistent clones!

in order to make the VSI VDI plugin to work with Horizon View, you will need to disable the message security mode (as the vm’s creation and communication about it is coming from an “external” source to Horizon.



VSI 6.2 Release Notes:


VSI 6.2 Download:


VSI 6.2 Product Guide:


Leave a comment

Filed under Uncategorized

VPLEX/RecoverPoint Integration with XtremIO – A customer case study


It’s not a secret that one of my favorite products within EMC is VPLEX, it’s a product that means different things for different people, some look at it as a DR solution while others loos at it in it’s true flash and glory meaning, a distributed cache that is virtualizing your underlying storage and provide an Active – Active site topology so for example, you can have site A in NY distributed  to site B in a remote location (up to 10ms latency is the supported configuration as of 07/2014)

“but isn’t the latency between the sites defied the point of using an AFA like EMC XtremIO” ??

nope, you can have you site A VM’s, DB’s et’c doing their thing in site A and in a n event of a site failure, leverage vSphere HA to restart them in the remote site


you can leverage the VPLEX RecoverPoint splitter replicating to a remote site in an async fashion, in fact I have just the customer that is doing this to tell you about..



The Tel-Aviv-based Clal Insurance is Israel’s leading insurance, pension, and financial  services group. The Clal Group holds a 19% share of the Israeli insurance market and  manages hundreds of billions of shekel-denominated assets. Haim Inger is the CTO at  Clal Data Systems, the IT company that provides all IT services to Clal Insurance.

Their Challenges

For Inger, the main storage-related challenge facing his organization was end-ofmonth reporting on the life insurance systems. Even with high-end storage in the data  center, these core applications needed better performance.
“The end-of-month jobs would start around midnight, and they would not end until  around 2:00 p.m. the following day,” he recalls. “The result was that almost 600 users  were unable to do their work for most of the day on the first of every month, and  customers were unable to get any service during that same period of time.”
This situation was costing Clal Insurance an estimated $900,000 per year in lost  business and productivity. Inger could see that the main bottleneck was storage, and  he was looking for a solution that would make it possible to finish these jobs much  earlier.

The Solution

The answer was an EMC XtremIO All-Flash Array. Clal Insurance started by deploying  a two X-Brick cluster (X-Bricks are the basic scale-out building blocks of XtremIO  arrays) and replicating them locally with EMC VPLEX to another two X-Brick cluster.
The local VPLEX replication of the X-Bricks was in addition to a full disaster recovery  (DR) backup system, using EMC RecoverPoint.
With this configuration, which provides both data availability and storage virtualization
for data mobility, Clal Insurance was able to dramatically reduce the time needed to  run the end-of-month reports. “Our goal was to finish the reports around 8:00 a.m.,
even though users typically arrive at work an hour earlier,” says Inger. “We actually  finished at 5:30 a.m.”
And the story gets better. “We have seen the same benefit in an even more
demanding scenario, which is end-of-year reporting,” Inger continues. “On a usual  basis for end of year, we would be shut down for 24 hours. Thanks to XtremIO, we  were able to open the systems at 6:20 a.m. on January first of this year. Even with  local replication using VPLEX, plus RecoverPoint backup to our DR site, the reporting  goals and online day-to-day work were not affected. We don’t see any impact at all.


Clal Insurance tried several ways to solve the end-of-month reporting issue before
deciding to deploy XtremIO. First they installed more flash drives in the existing
storage they had on the premises. This approach resulted in small gains — instead of  2:00 p.m., the jobs finished at around 12:00 p.m. — but it was not enough. Then they started looking at All-Flash Arrays, including XtremIO, and others.
“The reason we decided to go with XtremIO, even though it wasn’t yet generally
available, was the product’s ability to coexist with our current environment.” says Inger. “We already had VPLEX installed, and our existing DR configuration was based  on VPLEX and RecoverPoint. The only solution we saw that was really plug-and-play  was XtremIO.”
Clal Insurance evaluated two X-Bricks for almost six months before putting XtremIO into production. “We tested two things,” explains Inger. “First we tested the stability, to see that we didn’t get data corruption, downtime, or anything like that. XtremIO  was very stable, even during the early beta phase. The other thing we checked was performance. Depending on the application, we got from two to ten times better
performance with XtremIO compared to our existing storage environment.”
Specifically, the best average read time on the company’s existing high-end storage was about 5ms. Clal Insurance was looking for at least 3ms; they achieved 1.5ms with XtremIO.
Inger says XtremIO was the right choice. “From the moment we got the X-Bricks and installed them physically, it took us less than a week to move the first production database over, and then an additional two weeks to move the other four databases —a total of 14 terabytes,” he says. “And it just worked. I didn’t have to reinvent or rearchitect my DR system or anything else. With any other solution from IBM, NetApp,
or the All-Flash Array start-up vendors, I would have had to reinvent almost
everything about backup and DR. What’s more, the management is any storage guy’s dream. XtremIO is the simplest storage system I ever saw anywhere to manage.



Unlike other insurance companies, Clal Insurance is 100% virtualized. The company has more than 1,300 servers, all using the latest VMware solutions. XtremIO fits right  in. “The benefit of XtremIO for a virtualized environment lies in its Inline Data  Reduction capabilities.” says Inger. “Our database servers previously consumed about  120TB of Fibre Channel disk. With XtremIO, we consume only about 40TB of SSD
because of the data reduction ratio. That is a real savings of money and footprint, with much better performance than we had before.”
Inger purchased several additional X-Bricks recently. “All of our database virtual  machines are going to reside on XtremIO,” he says.
In summary, XtremIO is used for two main purposes at Clal Insurance: The first is to  accelerate the day-to-day use of applications and to super-accelerate the nightly and  end-of-month jobs of the life insurance applications. The second is to accelerate the deployment of the company’s virtual servers with as small a footprint as possible on  the storage side.
“The greatest benefit we have realized from our XtremIO system is the ability to serve  our employees and our insurance brokers as fast as possible, and with no downtime,  for end-of-month batch processing times,” says Inger. “This project really shows that  investing in infrastructure can bring great ROI to our company while improving the  overall service we give our customers. And we have found XtremIO to be totally reliable, with zero problems.”
According to Inger, XtremIO has a bright future at Clal Insurance. “I plan to move all of the mission-critical databases and most of the mission-critical applications to  XtremIO over the next 12 months,” he concludes. “In my opinion, everything will go there within three to four years.”

below you can also see a demo of VPLEX + the RecoverPoint splitter on XtremIO

Leave a comment

Filed under Uncategorized