Straight Forward Convergence with Scale

I have to be honest here, I’d heard of Scale Computing before but never really paid too much attention to them. That is, until I got to see them present at Virtualization Field Day 5 in Boston Massachusetts this year.

All travel expenses and incidentals were paid for by Gestalt IT to attend Virtualization Field Day 5. This was the only compensation given and did not influence the content of this article.

Don’t forget about these guys if you’ve bought into the Hyper-Converged Infrastructure idea. Scale Computing targets small and medium businesses specifically, due to their low cost solution. The Scale nodes come in three models currently and the best part is that you don’t need to talk to a sales rep. to find out how much they list for.

HC1000 = $25,499

HC2000 = $37,499

HC4000 = $67,499

These prices come with a few caveats though. The existing version of the Scale Computing platform only supports a max of 8 nodes. You could have multiple groups of nodes but you would have to manage each set of 8 independently from one another as of right now.  Also, if you really want to run Hyper-V or vSphere as your virtualization solution, you’re out of luck. The Scale solution leverages KVM for a hypervisor and this is likely whey the prices is so competitive. Also, there is not a lot of customization available for how availability such as number of failed nodes the cluster tolerates or making sure that VMs are housed on separate nodes (anti-affinity).

The hyper-converged infrastructure solution touts an easy to administer HTML5 GUI that you could run from a smartphone if necessary. The clustering setup takes a pretty complicated set of Linux instructions and places them into an easy to configure GUI. The solution also provides asynchronous remote replication if you are in the need for a disaster recovery solution.

Design Decisions

If you’re in the market for a new solution, take a second and consider your requirements though. Do you need any of the feature that are only available with vSphere or Hyper-V? Do you need a single cluster with more than 8 nodes? Is your budget constrained so that you can’t afford a more expensive play? Do you need to specify specific availability requirements, or do you want the system to be available after losing a node with minimal downtime and don’t want to think about messing with the configurations?

Scale Computing definitely has a place in the industry, even if it’s a small niche. They currently support over 1100 customers and are doing so with a support staff of eight. “The entire company is only sixty-five people” CTO Jason Collier noted during the VFD5 session. It is obviously a point of pride with the company that they are doing so much with so little.

Summary

I won’t sit here and tell you that I think this is an enterprise ready solution but for a small to medium business with very few IT Administrators, this may be a pretty good play. The ability to purchase your compute, hypervisor, and storage all in one shot and getting some enterprise capabilities (replications, clustering, etc.) is pretty nice. Combine this with a pretty nifty price tag and easy setup and you’re well on your way. I’ll also throw out there that it seems like a risk that Scale Computing only has eight tech support engineers, but it’s also pretty likely that you’ll have a good relationship with them. If you call for support more than a few times, you might know these guys by name!