This podcast is sponsored by:

Cirba-logoIn this episode I’m with Andrew Hillier (CTO and co-founder of Cirba) and we talk about:

– The company
– Product architecture
– Use cases

Here the transcript of the show

Enrico: Welcome to our new episode of Juku Beats. Today I’m with the Andrew Hillier, CTO and co-founder of Cirba. Hi, Andrew.

Andrew: Hi, Enrico. How are you today?

Enrico: I am feeling good. First question is about you and about your company. Can you tell us about Cirba?

Andrew: Sure. As you mentioned, I’m the CTO and Co-founder, the company name is Cirba. Really in a nutshell what we do is analytics that helps you align you supply and demand in your data center, in your virtual environment, in your cloud environment. We take a very comprehensive look at the work load patterns, the requirements of those work loads, the hosting infrastructures that are there and you can think of this, we do one thing really well. We figure out where to put things. We figure out how to combine workloads to get the maximum efficiency, we figure out where workloads should run cross-hosting environments, across different cloud technologies, like VM or OpenStack. That’s really what it’s all about. It’s all about driving up high-efficiency, driving out risk, and where to put densification. Creating high density so that you’re [inaudible 00:01:10] cost for hosting work loads goes down.

Enrico: In fact, I just had a look at your website and [Cirba 00:01:19] seems to have a very complete solution for utilizing environment. Ranging from control to [inaudible 00:01:25]. I know there is a lot under the hood. Can you describe the different components involved in all of this and how they work together to achieve a [inaudible 00:01:38].

Andrew: Sure, sure, absolutely. Under the covers, like I said, is this analytics and Unix policy-based analysis and we see a lot of organizations really start to focus on policy because they realize they need to quantify on how their environments operate. [inaudible 00:01:54] there’s several ways that it’s used. It all stems from the ability to figure out where to put things but there’s kind of different use cases we’re in there. We have a control console. It’s basically automated analysis that looks at how the VMs map onto the physical servers and the cluster in the host aggregates in an operating environment to make sure that they are [awfully 00:02:15] placed. At the lowest level, we do very, very specific VM to Host placement. It’s like playing a big game of Tetris. We look at the workload patterns and say, “That one’s busy in the morning, that’s busy at night.” Or “That’s CPU intensive, another one’s IO intensive.” We look at it all in combination and say, “This is how we can combine them together.” Typically you’ll get about 48% increase in density by doing that. It’s very dramatic.

At the same time, we also look at how to place the VMs so that we minimize the software licensing costs. For example, for Windows operating systems or for data bases, by combining them cleverly on fewer servers, we see on average a 55% savings, for example, on Windows operating system licensing. That’s the densification I was talking about. Where really by being a little more clever about where you place your workloads in the environment you get a big bump in efficiency and your risk goes down at the same time.

That’s one use case, that’s what we call controlled. If I step back from that for a second and look across my entire enterprise, most organizations have multiple hosting environments, multiple technologies. They might have dozens of VMware environments, clusters, [asomunix 00:03:25]. We have a routing and reservation system that looks at inbound demands and says, “Where should I put them across all my environments?” It’s almost like for IT. You can say this inbound application, it needs an Oracle license, it has customer data, it needs to run in this country. We can look at that and then match it up with all the infrastructure because of our detailed control analysis of all the infrastructure. We know which environments have the right amount of resources, they have the right kind of resources.

That’s the routing and reservation. We can make a very high-level decision about where something should run. If it’s not going into the environment today, we also reserve the capacity. It’s not just writing down a quota, we actually model that capacity into the environment so you’re sure to have it in two weeks or a month or whenever you need it.

Those two pieces are really the two main use cases. One is, we call it kind of supply side control, and driving efficiency there. Then the demand management side where we’re looking at the [inaudible 00:04:28] existing apps and figuring out at a high level where they should be placed. That which case is becoming more and more interesting because we’re seeing people running, for example, multiple cloud technologies, maybe OpenStack and VMware in parallel. This can also make the decision between those technologies and say, “That particular workload should go into an OpenStack environment or VMware based on its exact requirements.”

Enrico: Sounds real interesting. You already touched what I wanted to ask you about use cases. Can you spend a few words about that customers that are using this solution? Who is deploying Cirba and [inaudible 00:05:15]?

Andrew: Sure. We have some very large customers, we have some small customers. The nice thing about this solution is it gives you that savings in large scale or in small scale, it’s kind of the same densification message. We have some very large banks, insurance companies, logistics companies that use our product. In most cases it scales out to over 100,000 physical and virtual systems. Those are great use cases because that scale of course driving [inaudible 00:05:44] has huge benefits. If you could be more clever about how you place the workloads, you could save millions of dollars in a hurry. You can avoid buying hardware. A lot of these organizations have very complex compliance requirements. They might have rules on avoiding risk by not concentrating things together or a certain compliance rules for certain kinds of business. All of that is a great use case for us because it requires the analytics to figure out where to put things. We even had customers that are getting pretty advanced into SDN, software defined networking. Again, that creates an even bigger optimization potential, but it’s also a bigger challenge because it allows workloads to be even more mobile. You can play an even bigger game of Tetris. The rewards are higher but you really need analysis to do that because it’s even harder to figure out.

We have some really interesting use cases kind in larger organizations and the more forward-thinking organizations. We also have customers that are much smaller. We go to market also through customer or sort of partners that include our product in their offerings. We have customers right down to 250 VMs that are running in a cloud that’s being managed by our product because it’s being offered by one of our partners. We kind of have a full range of customers from very large organizations controlling and running to smaller ones that are benefiting from it because it’s kind of plumbed into the cloud solution that they’re using.

Enrico: Thank you very much for all this explanation. Where we can find you on the web or on Twitter?

Andrew: Our website is That’s C-I-R-B-A. Our Twitter handle is @Cirba, capital C-i-r-b-a.

Enrico: Great. Thank you very much for your time, Andrew.

Andrew: Thanks, Enrico.

Enrico: Bye.

Andrew: Bye-bye.