Contributed article by Andrew Hillier, co-founder and CTO, CiRBA

Software Efficiency, Multi-Hypervisor and Software Defined Challenges to be seen in 2014

In 2013, organizations turned their attention to the problem of over-provisioning, and explored the levers that enabled them to move to the next level of operational maturity and efficiency within virtualized infrastructures, and of course, the cloud.

In 2014, the drive to efficiency will continue, but will not just focus on reducing hardware spend.  Significant savings will also be had through increasing the density of expensive operating system and application licenses.  Many organizations have already adopted per-host or per-processor licensing models for this very reason, but have not yet optimized the density of the VMs to realize the savings.  Some forward-thinking organizations have already placed a big focus on this, but the coming year will see this happen on a much broader scale as it becomes clear that a lot of money can be saved by simply moving VMs around.

This software efficiency theme will also extend to the hypervisors themselves, which have long been a sore point from a cost perspective.  Many organizations will seek to reduce the unit cost of hosting workloads, and at the same time avoid vendor lock-in, by considering different hypervisor alternatives.  Hyper-V environments will become more commonplace, and KVM will start to gain traction as it rides in on the coat tail of cloud technologies like OpenStack.  As with any adoption cycle, these trends will start in dev/test environments, but will have an accelerated path to production as the ecosystems build out around them and management vendors throw more weight behind them.  Those organizations that remain flexible and utilize the automation tools that optimize workload placements will continue to see efficiency gains.

Additionally in 2014, as more and more infrastructure components become "software-defined," organizations will realize that the more degrees of freedom there are to define things through software, the more difficult it becomes to figure out how to define them.  We have had a glimpse of this already with virtualization, which is really another name for software-defined servers.  Although it created the ability to place VMs on different servers and flexibly define their resource allocations, it ended up causing a bit of a mess as VMs were inevitably put in the wrong places and made the wrong sizes.  In recent years management software has emerged to control this flexibility, and by analyzing all of the operational metrics and constraints it became possible to optimize placements and allocations, effectively driving up efficiency and reducing operational risk.  The need for this kind of approach will expand to a broader scale, and from it will emerge software to define the software-defined data center.

##

About the Author