The Software Defined Data Center (SDDC) is considered by many to be the final step in the evolution of virtualization and cloud computing where all IT infrastructure (servers, storage, and networking) is virtualized and applications are delivered as a service internally or externally. Defined by IT people to mitigate potential IT problems, this model however leaves the data center’s critical physical and mechanical infrastructure out of the equation.
As a result, application outages happen now more than ever, due to power and cooling issues. Therefore a complete “Software Defined Data Center” also has to match the scope of the original physical data center, including not only IT servers, networking and storage but also power, cooling, and the building itself. For ultimate reliability, the application also needs to be abstracted from the data center power and cooling systems.
With extreme weather, blackouts, brownouts, and other catastrophic events are happening more frequently, and due to 24×7 application requirements the promise of SDDC is incomplete without accounting for the reliability of power delivery to and within data centers. “Software Defined Power” is a mechanism to shift an application to the data center with the most reliable and cost efficient power source at any given time—within the limits of application service level guarantees—and is therefore hugely valuable in conquering the challenges of application reliability, as well as saving energy costs.
Clemens Pfeiffer, CTO of Power Assure, Inc. recommends the following to ensure the proper planning of a complete SDDC environment.
- Leverage software defined servers, software defined networking, and software defined storage solutions to free up applications from physical IT equipment.
- Add a Software Defined Cooling solution to allow for dynamic adjustments of cooling capacity based on the actual heat output of IT equipment under variable load conditions.
- Add a Software Defined Power solution that can migrate applications from one data center to another and provides power grid integration to intelligently determine the most reliable configuration for data centers at any given time.
- Think of the software-defined data center as a pool of buildings, IT, and cooling resources that can be used for applications as needed depending on application demand, power cost and availability, weather pattern, and resource availability.
- Using Software Defined Power as part of the Software Defined Data Center is the path to Ultimate Reliability.
According to Pfeiffer, “To achieve ultimate reliability, IT and facilities components must be integrated, managed centrally, and with most (if not all) of the underlying complexity abstracted away to a software-control level where balancing applications across data centers is fully automated and all components are dynamically adjusted based on variable application load levels. When done correctly, this will not only increase reliability but also cut operating costs by over 50% and allow data center operators to benefit from energy market incentives and pricing spread between the different locations. Forecast models as part of Software Defined Power will help adjust operating schedules according to the latest energy market conditions to maximize such incentives while minimizing the risk for the application.”