That old adage, it’s not what you’ve got but how you use it, takes on a particular significance in the context of the virtualized data centre.
While virtualization’s intrinsic ability to radically streamline IT infrastructure, drive efficiencies and boost speed and agility, has cemented a certain game-changing status, for all the hype and capabilities, the full potential remains routinely under-explored.
It’s a scenario borne out by Gartner research which found that despite 60% of workloads now adopting this approach, just 25% of the available processing power is actually used, some way off the 55 to 60% target that needs to be hit to achieve the true economies promised.
Fuelling this issue are assumptions that deployment alone will be enough for the VMware tools to work their magic, an oversight compounded by the common default approach of adding new physical servers to drive performance rather than fully utilising the existing ones.
Remedying this oversight, requires a shift in focus to adopt a more strategic management of what is already in place, essentially making more of what you’ve got. Only by understanding how the workloads or virtual machines (VMs) should be combined and placed into the underlying infrastructure to make the best of available resources, will see the dividends reaped. And pivotal to this is the role of predictive analytics, offering a deeper level of insight to best drive VM placements to reduce their movement and lower software costs.
It’s an approach that is emerging as a key enabler in large-scale software-defined infrastructure environments and reclaiming the focus from the often unrealistic expectations around the role of real time load balancers to effectively assess and respond to all requirements in what are highly complex environments.
Indeed, the issues are further compounded by technologies such as Software Defined Networking which frees the virtual environments from the traditional confines of the physical network boundaries, enabling virtual machines to roam freely across the infrastructure. The ability to understand where to place workloads to avoid risk and increase efficiency with a forward-looking view of demand cements the critical role of predictive analytics. For leading software-defined infrastructure control provider, Cirba, harnessing this power is seeing the Canadian-headquartered company emerge as a true technology differentiator in the market, driving next generation hosting with analytics that scientifically balances the infrastructure supply with application demand to remove risk and increase efficiency.
Underpinned by predictive analytics which ensures optimization is automated and ongoing, the demand-driven solution seeks to optimise the use of all key compute, storage, network and software resources in the environment.
Ongoing analysis ensures the approach takes the guess work out of how to size and place workloads, and eliminates stranded capacity by balancing VMs across an enterprise based on observed patterns of usage and operational and business policies. Critically, the VM density in the environments is radically enhanced without compromising application and workload performance.
To get some idea of how this works in practice, it’s worth a brief revisit to the mid ‘80s when the world was hooked on the Tetris video game and the challenge of piecing together the blocks according to size, shape and orientation for best performance. For Cirba it’s a similar process; multi-dimensional analysis considers all technical and business constraints, while also looking at the workload patterns, personalities and profiles to optimally fit workloads together to make the best use of resources without introducing contention for resources. Cirba does this within a cluster and can also balance demand across an organizations many environments by determining the optimal hosting environment.
It’s an approach that is spawning some very impressive outcomes, achieving an average boost in VM density by some 48%, a 33% saving in hardware costs and over 50% in software licensing, all of which has a significant impact on overall ROI which those in the know have been quick to exploit.
A diverse range of customers are testament to the benefits, including several of the world’s largest banks, one of which turned to Cirba’s analytical prowess to curb the additional costs that would be incurred as a result of a license renewal and saved itself $82 million in the process over five years. Another customer realized a documented $24M in hardware savings, when building a new private cloud.
Marrying the transformative powers of virtualization with the forensic intensity of predictive data analysis, is proof that even the most game-changing technologies can use some help to reach their full potential.