Software-Defined Anything Takes Root

The concept of software-defined anything — or SDx — is still evolving, but the endgame is a powerful one: an infrastructure with the smarts to tune itself to the needs of the applications running on it.

Generally, it's agreed SDx entails leveraging software control layers to intelligently, holistically and automatically configure and manage physically abstracted and diverse computer, storage and network resources. That idea has a lot of power over the future of IT.

The roots of SDx lie in software-defined computing, "which we call [server] virtualization," says The Hackett Group Principal and Co-Founder Allan Frank. "When it comes to virtualization and all these SDx concepts, we're really talking about the holy grail of how we can start to take the entire infrastructure of an enterprise and allow it to operate effectively by itself. The infrastructure becomes smart enough to tune itself to the needs of the applications it's running without human intervention."

The ability to bring more intelligence to infrastructure management matters a great deal in many higher ed institutions, where IT often struggles to attend to increasing needs and expectations with fixed or declining resources.

Case Western Reserve University has been "working to create capacity in order to increase support in the areas of research, teaching and learning, which is the university's core mission," says Mike Kubit, Case Western's interim chief operating officer of Information Technology Services. "Finding ways to more effectively manage our infrastructure is critical to creating that capacity."

SDx in Action

While there was a time when a ­university's infrastructure served as a differentiator in attracting students, a robust infrastructure is a basic expectation today, Kubit says. That makes the promise SDx brings of "managing massive networks and data centers with fewer people" a very appealing one, he says.

Case Western has experimented with SDx in the form of software-­defined networking over the past 18 months. The team there found that the ability to intelligently control a network servicing a 155-acre campus and 150-plus buildings brings clear advantages.

"The benefits are not only the ability to control multiple devices and platforms, but also to gather data as far as how the network performs and whether we have resources allocated properly to keep users from having quality of service issues, and then adjust settings to tune performance from a central location," Kubit says.

From the operational perspective of maintaining routers and switches, "being able to build in some predictable behavior using data as far as usage and traffic patterns offers the promise of requiring less manual care and feeding on a daily basis," he says.

Off-Campus SDx

Frank says SDx will be important to cloud computing in all its forms: private, public and hybrid. It should also prove important to CIOs in the university setting — typically characterized by redundant and localized platforms and data centers, as well as distributed support of these technology silos — to join with managed service providers to virtualize and centralize institutions' transactional back-office functions, and later academic applications, in increasingly SDx-capable public clouds.

Third-party cloud partners can draw on smart technology such as SDx as it matures to further drive productivity improvements and reduce costs.

The Hackett Group's Global Benchmarking Practice Leader Anthony Snowball says universities mostly have lagged behind other sectors when it comes to outsourcing, but could start to realize the advantages that managed hosting can bring, from reducing staff requirements and internal IT costs to delivering greater efficiencies to their organizations.

"Innovative technology advancements such as self-service, the cloud or 'smart' technology like SDx may offer universities an opportunity to take a quantum leap forward from traditional and expensive IT service delivery models," he says. 

"Universities need to start focusing on managed services now versus continuing a doing-it-all-on-your-own computing model," Frank says.