Nexenta

The photo above shows a number of happy people from Nexenta and from our partner DataON and the top of a test rack of NexentaStor loaded onto two Intel servers plus an LSI 6Gbps SAS connected array providing a shared JBOD, with all single points of failure eliminated including power supplies, MPxIO paths, and of course the Intel based NexentaStor heads.    More information about the combined solution, which is undergoing final certification this week, can be found below.

DataON over the last year or so has become an important supplier of solutions to the Nexenta channel.  We have large active / active deployments in the US and Europe using their LSI SAS connected JBODs.  Customers have been extremely impressed with the quality of the hardware supplied — as they should as it is made by the same companies that supply leading legacy providers.

DataON is a great example of how the open storage ecosystem has gathered strength so that we all work together to provide total solutions that exceed the capabilities of legacy storage.  The reach of the open storage ecosystem extends from the storage hardware vendors themselves, in the US and around the world, to all interested server resellers in the world.   The size and depth of this channel is much larger than the traditional storage channel, which is also starting to adopt open storage as well.

Here is a little more about that solution, including draft test results of reads first and then writes showing 1600MB/S through put.  We’ll also be publishing results with deduplication enabled over the next few weeks.  And when we do so we’ll publish information about capacity utilized and so forth so you know the results are not entirely due to short stroking the drives and so forth.

Nonetheless, the headline, top line throughput is fast.  And IOPS from the same system will be quick as well.

write performance with different file sizes, draft

read performance with different file sizes, draft

Here is a marketing level schematic of what is creating these impressive initial test results:

These solutions are achieving solid 4 9 reliability in the field AND great performance.  Thanks to NexentaStor they have a number of capabilities lacking on legacy storage systems at any price - including the detection and correction of silent data corruption, in-line data deduplication for primary and replicated data (with 3.0) and the industry’s leading solution for the management of storage in virtualized environments.  Purchased through a local partner anywhere in the world you can save 60-80% or more versus high end legacy storage of similar performance and capacity and be assured of similar savings as you expand capacity in the years to come due to the absence of vendor lock-in.

Thank you to DataON for bringing us this excellent solution for final certification with LSI 6Gbps SAS.  And thank you testing team for sharing your preliminary results.

Any users of DataON and NexentaStor — please comment below to let us know what your experience has been like.

We have a good crowd attending next week’s training in Atlanta.  Learn more here.   Thank you for Rojoli for helping sponsor the event which will be held at the Atlanta Athletic Club.

I am very happy to announce that there will also be training April 29th-30th in Las Vegas the same time as Interop.  Learn more here. Prices increase April 1st — so register in March for savings.

What we’re already finding with these events is:

a) there is a gigantic amount to cover with ZFS and storage

b) primary storage deduplication & virtualized environments is of great interest (note — this is where we think our VMDC plays a huge role)

c) SSDs are also of particular interest.

So, we have taken the feedback and created training that focuses on SSDs, and again deduplication but that begins with a thorough view of ZFS.  Many have pointed out they are intending to visit Interop so we have scheduled the training for the end of April in Las Vegas.

As you’ll see when you look at the registration site, Richard Elling, who is a notable ZFS expert who also helps in various NexentaStor field engineering and deployments, is again leading the training.

Also, attendees will experience hands-on training in the use of SSDs and ZFS.  Attendees will have access to the DDR Drive X1, which is the first SSD solution built specifically to accelerate the ZFS ZIL.  Test results from NexentaStor plus the X1 show 300,000+ random read and 200,000+ random write IOPS.

This portion of the training will be led by the CTO of DDR Drive, Christopher George.  Chris is one of the world’s experts at using SSDs for improved price performance of storage.  We’ve been working with Chris for quite a while and our partners are already seeing the X1 as a real competitive advantage versus legacy storage arrays.

As a final benefit, attendees will be able to receive a Certified NexentaStor Engineer recognition.  This is particularly important to partners who need two CNEs on staff to be Nexenta Certified Partners.

I recently learned that of my Grandfathers had something to do with the use of Native Americans to secure US government communications during and after WWII.   Wikipedia has a nice write up here.

It occurs to me that the use of proprietary storage is a bit similar to the use of Navajos for secrecy.  What if companies stored their data on paper, still, AND instead of using their local language such as English they handed everything off to a Native American who transcribed everything into their language?  Over time, those enterprises would be very good customers of the Native Americans.

And if the Native Americans wanted to extend their business a bit they could even sell proprietary paper.

This metaphor is stretched to the point of breaking already…. But storage vendor lock in really is just that fundamental.  Enterprises have to schedule a several month process to get their data translated from the proprietary on-disk formats in which it is stored.  The data — and the customers — are truly ‘locked in.’  So the price of paper, sorry, disks from such a vendor is 10-20 times more than if the disks were bought from the disk supplier.

By comparison, an OpenSolaris user who wants to use NexentaStor to manage their data can simply install NexentaStor on the server connected to the drives with the ZFS data or they could even physically move the drives.   In a few hours you can switch your storage vendor for hundreds of TBs of storage.

The lack of vendor lock-in keeps Nexenta incredibly focused on where we add value, our software, our support, and the total hardware / software solutions sold by our partners .   We have no customer lock-in and can only continue to grow quickly by adding value to our product.

Any NexentaStor users reading this — did you add NexentaStor to manage existing ZFS data?  Or did you migrate away?  If so, please comment below, I’m sure others would be interested in your experience in avoiding vendor lock-in.

I am still at CeBIT where we have had many, many end users and partners interested in, for the first time, being able to run software plus industry standard hardware and arrive at a solution noticeably superior to legacy solutions. Much of the interest this week may be prompted by our announcement of the addition of primary storage deduplication to NexentaStor. While this and other improvements grab the headlines, inside of Nexenta we know that it is the “boring” things that win and retain customers.

We have committed ourselves at Nexenta to becoming world class at providing enterprise class storage solutions in partnership with  storage and server channel partners who deliver total solutions comprised of NexentaStor, industry standard hardware, and services.  Being successful at this model requires training a lot of partners on such tasks as how to discover customer requirements and how to explain exactly how NexentaStor and OpenStorage can best address these needs.  And, of course, it means building a solution that requires little in the way of support and that is easy to support when support is required.  While customers want great features, they fundamentally require accurate and insightful information when they are considering alternatives and also when they are running their business on NexentaStor.

And getting that information to end users, via partners, is not always simple.  For example, I just had a software company partnered with SAP stop by and ask if he could use NexentaStor in a HA cluster to provide seamless failover to the VMs.  He had been told that he cannot by a partner who is one of the most technically advanced partners we have.  In fact he can - IF he uses shared storage.  He cannot (yet) have a completely seamless failover if the storage is NOT shared.  This was the confusion.

Thankfully - due to lots of work by our marketing and solutions teams and especially the partners themselves who collaborate extensively - the above case is the exception and not the rule.   A critical mass has been reached and best practices about NexentaStor and ZFS are better shared and understood than ever before.  Of course events like the training March 16-18th in Atlanta, Georgia help as well.

What do you think?  How are we doing at sharing useful information.  And how would you like Nexenta to share information about capabilities and requirements?  What approaches have you seen work well?