Optimizing the data center environment – for density, efficiency and cloud – is a challenge a number of vendors are rising to, propelled by the popularity of anything to do with the software-defined data center (SDDC). But some of the techniques used to position data and infrastructure in the data center (and now between data centers) are not necessarily new. In many cases, it is just that a label had not yet been created for what was on the table.

FOCUS recently caught up with CiRBA, a company that has played in the data center analytics space – not to be confused with business intelligence or analytics – since it was founded in Ontario Canada in 1999. Its goal back then was to provide what it calls “automated capacity control” software for virtual and cloud infrastructure.

In London, SVP for Europe, the Middle East and Africa Ayman Gabarin explained how the technology has moved forward driven by the fast-changing environment and vendor and technology landscape operators now find they are dealing with. The challenge for CiRBA today is that with the advent of the term ‘software defined’ the company is crossing many boundaries – two of the most distinct being between capacity planning and IT service management. It does this by defining policy, managing the estate and providing assurances of service levels for capacity demand.

“We are kind of calling this automated capacity control, not management, the difference being that you manage your finances once a month but you control your cloud with our hands on the wheels,” Gabarin says.

What does Gabarin mean?
CiRBA has created a policy-based engine that can automatically control VM placements and predict future capacity requirements using software. To give you an idea of the basics of the technology, we can look at a deployment from 2008.

Australian-based digital service provider Hyro added CiRBA’s data center intelligence software to its consolidation and virtualization analytic software tools offered to enterprise clients. It did this to analyze client environments and plan consolidation and virtualization initiatives. Back around this time, CiRBA’s CTO Andrew Hillier described CiRBA as being the “brain” of the data center. For Hyro, it was another way of offering that little bit extra to the customer. Pushing forward to 2014, it seems differentiators such as these are being dreamt up by service providers around the world facing the world of cloud, where the underlying infrastructure is only as good as the services a provider can offer on top.

CiRBA  “understands” what is in a virtualized or cloud environment from a server, storage and network perspective and where the best place for application workloads is.  “It can understand the SLA, the storage requirements and all other metrics, including regulatory constraints on the business,” Gabarin says. “In the past people managed the disconnect of their portal for requests, and back end of operations, using spreadsheets."

CiRBA does not touch the data itself, but it does have control over where it moves. “The way we do that is we have the concept of each piece of infrastructure, such as servers - physical and virtual - as well as the application and the rules that applies to each. We then provide unlimited levels of service and rules that you can define in each product, so when we analyze an environment we can tell you if your apps are sitting on the wrong infrastructure from a data compliance point of view, or how you should rebalance your environment on a daily basis to achieve your goals for energy efficiency, density, cost and so on,” Gabarin says.

A Reservation Console (launched this year) receives requests for new workloads such as from self-service portals, and together with the observed workload requirements and trending, provides for a comprehensive model of future demand. This allows for future capacity planning.  The console was released as part of CiRBA’s latest version -  Version 8 – which supports VMware ESX, IBM PowerVM, Red Had Enterprise Virtualization and now Microsoft’s HyperV.

“The latest version is something we have been working on for a while. It covers the demand side of managing the data center enterprise. We have always had capabilities to manage what you have got to run – the pipeline of what is coming in and the demand from users. With the reservation console we can plan into the future and capture what is happening across the business, and with HyperV it means we can do this without customers being locked into any specific technology. You can define the options for the living service you want to consider. It might be internal cloud, or some other environment not part of your primary cloud sitting inside another company. You can now have the same view of this other environment, even if it is Amazon or IBM. You can now say ‘in 60 days from now I will run out of capacity, so I want to address that in 30 days’,” Gabarin says.

CiRBA allows you to drag infrastructure, literally, from a data center into a bucket with Amazon, IBM, HP, and say how many instances you want and figure out what the cost is. “You can then see how much free capacity you will have for future demand. You can also shift things between your own data centers. A lot of organisations mix things together with different SLA requirements, and CiRBA can even analyse where the best place for these to go to reduce cost and meet their requirements. Essentially you are cleaning up the clutter and making sure things in your data centers are in the right place,” Gabarin says.

Where CiRBA fits in to the SDDC
Gabarin claims the capabilities of CiRBA’s software sits at the heart of the software-defined data center. “We are now integrated with a lot of the orchestration layers and engines because none of them have the ability in terms of being able to make an intelligent decision on where to put things,” Gabarin says. “With many software defined options today, if something things it has capacity the software will just move it there as a random decision. It doesn’t necessarily have the concept of service level or suitability, or even the availability of sufficient resources for capacity from a CPU or other perspective, or the reservations of these for future capacity.”

In CiRBA’s view, millions of pounds of wasted infrastructure lays idle in today’s data centers to meet the demand fluctuations of tomorrow. “We can capture that demand in advance and plan for it way ahead to avoid this,” Gabarin says.

In future, Gabarin says he believes CiRBA’s software could plug into data center M&E management tools to further drive efficiencies with power and cooling.

One area where the technology has been driving down costs today, however, has been with software licencing. Many companies currently pay for licencing across their entire server estate “because they have no ability to understand where the Oracle data base, for example, is sitting on their servers”. “And if you want to use cloud you could have to pay for even more as the server estate expands,” Gabarin says.

“One of the things that we have done, because we can know where everything is and control this, is use analytics to drive where things are going based on rules. Customers can now ask CiRBA to analyse the environment and tell them how they should move things around to optimise density and relieve operational problems. A rule could be to class and condense all of the Oracle database on to the minimal infrastructure it requires, so then the CiRBA user can go to their friendly software provider and show them exactly where everything is deployed and reduce their software licencing costs. We have had customers realise significant savings literally over a weekend doing this.”