Software definition began with server virtualization and helped avoid a zombie apocalypse of dead servers; moved to the network with the emergence of the Internet of Things (IOT); and has now also reached the storage level of computing.
All these software-defined components mean that the data center infrastructure must catch up — beyond just physically scaling — to be equally agile.
Virtualization enabled consolidation — going from, for instance, 10 servers running at 10 percent, to 1 server at 80 percent; it freed up space and birthed a new level of flexibility. In doing so, it also created a gap in the traditional data center. Automation on the IT side calls for a software-defined data center (SDDC) to bridge that gap.
Software can now move loads dynamically between servers and even between multiple data centers depending on demand. For example, in retail, during the holidays, the amount of shopper data being captured is exponential. Plus, now the data is coming from many more points — i.e., sensors, smartphones, tablets, wearables, etc.
As loads fluctuate, the infrastructure — power and cooling needs — should match the heat output of the IT equipment. A sleeping server requires little energy and produces little heat, but at 2 am when a batch file starts running, it wakes up thirsty for power and may be as hot an oven.
A traditional data center wasn’t designed to dynamically respond to these erratic changes in power and temperature requirements. Many were overbuilt in an attempt address these possible situations at the expense of efficiency. But with an SDDC, power reacts effectively and efficiently through monitoring and automation via DCIM.
An intelligent, reactive infrastructure helps achieve resiliency, provides speed and reduces cost. This is critical for today’s pace of business.
Gone are the days over building and over provisioning and falling into the virtualization gap. A new day where IT and the physical layer are linked is upon us.
This is explored in a recent post by Henrik Leerberg, director of marketing and strategy for data center managed service & software, Schneider Electric, where he answers the question: “Is the Definition of the Software-defined Data Center Ambitious Enough?”