We received several interesting questions from an inquisitive Express News reader this week.  I thought they were fair questions, so I’d like to share pieces of our dialog (paraphrased and edited for clarity):

1. How is the Z2 Systems approach to providing a “virtualization platform” different than any other high-end server with a virtualization hypervisor installed?
2. How are your servers different from other hardware solutions offered by major server vendors?
3. What virtualization platform will be used?
4. What is your plan for supporting the hardware and software?

Great questions.  I’ll answer them one at a time, but first let me set the stage as to how we conceptualize the problem facing many firms looking to adopt virtualization technology:

Small and mid-sized businesses (or department/line of business requirements of larger companies) that implement virtualization solutions tend to be dissatisfied with the end result of their “lower end” implementations.  Or said another way, most of the low to mid-range virtualization implementations rarely deliver more than a modest consolidation of hardware – which is certainly an important benefit of virtualization, but not the only business benefit that a well designed virtualization infrastructure can provide (think reliability, disaster recovery, rapid test & evaluation, etc). 

Many virtualization implementations never get past the first step of reducing capital outlay costs.  They largely miss out on reducing other types of ongoing O&M costs, and in many cases they totally miss out on the ability to leverage virtualization to improve overall IT operational/support practices and/or serve as a foundation for better disaster recovery solutions.  This inability to optimize and leverage virtualization to the fullest extent can generally be traced to a single Achilles heel in their implementation approach – storage. 

Which leads us back to the reader’s questions:

How is our approach different?

We are essentially integrating advanced storage as a key component of the overall virtualization solution.  As firms recognize the operational and technical limitations of implementing virtualization servers on hardware with locally attached storage, and recoil from the sticker shock of high-end SAN implementations, many attempt to buy a NAS and configure their virtual machine hosts to access shared storage over the network.

These firms quickly find that while commodity NAS devices are great for storing large amounts of near-line data, they are absolutely horrible in environments with requirements for multiple, simultaneous, high transaction, and low latency I/O (the exact kind of data transfer requirements a virtualization infrastructure needs, by the way). 

Additionally, advances in storage back-end technology have resulted in the ability to do machine level point in time snapshots, cloning of any point in time snapshot, and even roll-back of a virtual machine to any arbitrary point in time.  Admittedly, many of these features – to one degree or another – have largely been part of NetApp and EMC solutions for quite some time.  Every hypervisor vendor and enterprise class best practice guide will tell you to implement your virtualization infrastructure against a high-end storage fabric if you want to really get the most benefit from virtualizing your infrastructure.  But those vendor storage solutions are not viable for most businesses, as they generally start at $50k-$75k and go up (quickly) from there. 
 
So you’re left with a choice: low-end commodity based hardware solution with only some of the benefits of virtualization (but none of the key operational ones) or high-end, expensive implementations (that most IT budgets couldn’t dream of affording).  We offer an alternative that doesn’t make companies compromise.  By building a “purpose-built” storage appliance and developing a tailored software application that provides for orchestration and management of the virtual environment, we have delivered (at a significantly reduced price) the same operational features/characteristics of those “best practice” infrastructures that cost upwards of $50k (and usually more than $100k).

In many cases these firms step back from virtualization all together, or scale back their expectations to meet their hardware budget.  We don’t believe that today’s technology requires such a compromise.

How is our hardware different?

We have integrated the storage and virtualization hardware in such a way as to ensure performance.  As an example, if you take a basic commodity server from Dell or HP, provision it with a high-end RAID card and enterprise level storage drives and slap a virtualization layer (hypervisor) on it (VMWare, XenServer, Hyper-V, etc), you will have a platform that can simply host multiple guest virtual machines.  Ok, cool.  It may do a reasonable job at it, but what about optimizing your storage layer as part of your overall solution? 

When your virtual machines have to compete for CPU, I/O and RAM with the underlying disk/storage subsystem, you can’t scale.  Likewise, when your virtual machines are all on a local drive (array) managed by the same operating system as your virtualization host, you generally can’t leverage fail-over, live migration, or offline backups.  Our Z2 Systems R3000x “Twin Server” really personifies this concept, as we pair our purpose-built storage platform with a very capable virtualization node in the same hardware platform.  The Z2 Systems R3000 is the same fundamental storage server (better hardware specs and expansion capability), but is tailored at larger environments or implementations where the client has already purchased virtualization hardware, or has larger overall requirements. 

Keep in mind that our software orchestration layer is what allows all of this to work seamlessly.  While we’re selling a tailored hardware solution, what we’re really selling is a vision of how to implement virtualization in a way that delivers better overall return on investment for a wider range of companies, and the software tools to make that a reality.

What virtualization platform do we use?

Citrix XenServer™ is currently our platform of choice.  We are evaluating support of HyperV and VMWare, but believe that the Citrix “Enterprise-class, Cloud-Proven” hypervisor offers the best platform for now, and into the foreseeable future.

As an early XenSource (acquired by Citrix) Partner, we have a significant experience and insight into the Xen Hypervisor and believe that the Citrix supported XenServer offers the best technical solution currently available – and we’re in good company, as more than 5,000 enterprise datacenters worldwide have already chosen Citrix XenServer and implemented it instead of alternative solutions. 

On 23 February, Citrix announced that the XenServer Enterprise version would be available FREE OF CHARGE(!!) for any user in production environments.  This has clearly changed the game with respect to overall cost – but in the same breath, they are receiving accolades for performance, and have been named the “Porsche of hypervisors” by one expert, (full Virtualization Review article available online).

What is our support plan?

Support comes in many flavors.  We provide hardware support through a nationwide third party OEM provider that builds the equipment to specifications under ISO 9001:2000-certified processes and provides 3 year Next Business Day onsite warranty support.  The hardware warranty support you receive is the same – in many cases better – support you would receive with the large “Name Brands”. 

Our team of software engineers and support technicians provides support (and ongoing upgrades) for our orchestration layer (free through our online support forums, or through annual support agreements with enhanced SLAs).  We have a robust product roadmap, and will be releasing some exciting new features in the next couple of months.

Our IT Solutions team also provides support to the virtualization software or even the OS and applications via traditional on-call support contracts or monthly managed service agreements.  We are available for engineering, design, implementation, and on-going support.

Wrap-Up

So, as it turns out, during our back-and-forth dialog with the (initially) skeptical reader, it was disclosed that he was a VMWare Engineer for over three years.  We didn’t convert him from VMWare to Citrix XenServer (yet!), but he did say, “I think that what you bring to the client will be a good value and the end-user will be able to realize a true ROI.”  Encouraging words.  What say you?  Please drop us a comment below: