Virtualization Technology News and Information
VMblog's Expert Interviews: Jason Collier of Scale Computing Talks Hyperconvergence in 2016

With hyperconvergence proving to be a hot topic throughout 2015, and the technology expected to become even hotter in 2016, I reached out to the experts at Scale Computing to learn more.  And luckily, I had the chance to speak with its founder, Jason Collier, to get the scoop. 

Scale Computing integrates storage, servers, and virtualization software into an all-in-one appliance based system using industry standard components. It is called the HC3 appliance which has a unified management capability driven by its patented HyperCore Software. 

VMblog: The hyperconverged market is a hot space right now.  Can you tell me, how does Scale Computing define hyperconverged?

Jason Collier: Funny piece of history is we were in the room when the term hyperconverged was coined by analyst Arun Teneja.  The reasoning for the term was to express when you combine converged hardware components of servers, storage, and networking with a hypervisor (hence "hypervisor converged" or "hyperconverged").

VMblog: Explain if you would, what's driving the hyperconverged market (or increasing opportunities)?

Collier: For us the 2 big drivers have been simplicity and cost.  The standard 3-2-1 (servers, switches, SAN's) architectures are just too complex and too costly for a lot of SMB's.

VMblog: As the market grows, where are you seeing the most traction right now?  And why is that?

Collier: SMB's across all of the major verticals.  We define SMB's in a pretty interesting way.  We define it as 1-5 man IT shops where the IT department isn't split up into SAN experts, virtualization experts, and network guru's.  I think the reason we have such traction comes back to simplicity being one of those key drivers, and that is a big reason for our success in that market.

VMblog: What does Hyperconverged Infrastructure mean for the SAN and server administrator?

Collier: Quite honestly you better shine up your resume! All kidding aside that SMB market that we build our systems for doesn't have those dedicated silos in the IT department anyway.  Everyone has to be the jack of all trades IT guy who does everything from configuring email on the CEO's new iPhone to configuring LUN's on the SAN.  Reality is the SAN and server guys will be just fine as they are primarily working in bigger enterprise organizations.

VMblog: That leads me to my next question.  How complicated is your solution to get up and running?  Are special skills required?

Collier: It's about as easy as setting up a linksys router.  You give the nodes an IP address, initialize the cluster, and you are up and going.  We did a live event with the folks over at Spiceworks where we un-boxed the gear, racked it, configured it, and we're presenting the interface out to a live audience in under 30 minutes.  The attendees were then able to jump on the cluster and start creating virtual machines.

VMblog: I've been asked by readers to explain the difference between converged and hyperconverged.  As experts in this field, can you help me out?

Collier: This is an analogy that I love to use.  Converged is an iPhone without iOS.  Hyperconverged is the whole package hardware and software. 

VMblog: What are some of the benefits a company could expect by moving to a hyperconverged environment?

Collier: There are quite a number of benefits.  First off you really need to see these things to realize just how simple they really are.  That simplicity means you don't have to spend a ton of time setting up, training folks, and managing these systems.  One of our systems engineers went home one night and shot a video of his daughter deploying a Windows 2k8 server.  It took her 45 seconds and she is 4 years old.  I love analogies so here is another one.  Instead of focusing on being the mechanic of the infrastructure, the admin can focus on being the driver and moving the company from point A to point B instead of just trying to keep the damn thing running. 

Second, you have one back to pat and one throat to choke if you have a support issue.  No longer do you have to sit in one vendor phone queue only to be transferred to another vendor phone queue to get your problem solved.

Finally, there is a significant cost benefit as well.  One advantage to Scale is that you don't have to license a separate hypervisor, or go out and buy a SAN.  This makes the initial and ongoing costs a fraction of what they could be for a 3-2-1 style of architecture.

VMblog: If a company isn't fully virtualized yet, is this a good solution for them?  And can this technology help get them over the hump to become fully virtualized?

Collier: Absolutely!!!  In fact it's the perfect time to come over to virtualization, because you don't have to spend the time to learn all of the complexities of VMware and the other virtualization solutions in the market space.  We have tools that can migrate physical workload to virtual workload (P2V) and also from other hypervisors (V2V).

VMblog: What separates Scale Computing from the hyperconverged pack?

Collier: Since day 1 we have controlled the entire stack from top to bottom, meaning we have our own OS, our own distributed storage stack, and our own hypervisor.  This means there are no other components you have to license.  There are basically 2 thoughts on how to accomplish hyperconverged.  You are either a stack owner, or a stack dependent.  A stack dependent runs as a VSA (Virtual Storage Appliance) on top of a hypervisor, basically they are a server-san.  Stack owners control the entire stack including the hypervisor, the hypervisor does not control them.  Scale is a stack owner.

VMblog: You guys are built on top of KVM rather than VMware.  What made you decide on KVM as the platform of choice?  And does it provide you with any advantages?

Collier: Back to the last question we use KVM, but it is built on top of us, and not the other way around.  With KVM we have complete access to the source code, and when you have access to the code, you are much more in control of your own destiny as a company.  Relying on a 3rd party company, who may not share your company goals and objectives as theirs, is a very dangerous place to put your trust.

VMblog: How well do hyperconverged environments handle high performance applications?  Any issues here to be aware of?

Collier: I'll give you a classic IT vendor answer here.  "IT DEPENDS".  Reality is that it depends on many factors, most of which evolve around how the application itself is architected.  Applications built for scale out style of environments tend to do well.  A single threaded database developed 30 years ago, well not so much, at least not yet. 

VMblog: Can you tell us, what's next for Scale Computing?

Collier: 2016 is going to be a very exciting year.  We are planning on some exciting product announcements in the first half of the year.


Well, sounds like a fun "stay tuned" from Scale Computing... a company with more exciting things on the way!  Thanks again to Jason Collier for taking time out to speak with VMblog.

Published Tuesday, January 05, 2016 8:01 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<January 2016>