When the green IT buzz first broke it was all about how enterprise data centers could be made more efficient. We heard about better building design, standards of measurement, improved hardware and software functionality with better automation for management. (The Green Grid continues some wonderful work here.)
Then came cloud. Here were the means to improve IT function, simplify operations and reduce costs. What followed was the debate around public, private and hybrid clouds. While that’s settled down, there are still many companies undecided on their direction, and a debate continues around the energy aspects of cloud computing.
Take the three most likely options – traditional IT operations, cloud or hybrid cloud. All hold promise for greater business efficiencies including positive energy and sustainability potential. Businesses need to understand the options and choose a best match for their circumstance and self-interest. Pick internal operations, cloud or some combination – what is the data-value return on investment? What are the bottom line impacts of IT efficiency per square foot or per yen/pound/euro/dollar of IT expense?
In 2007, when a growing data center energy crisis first spawned the rush to green IT, we were still operating in a pre-cloud model. Every enterprise had a data center or two and often maintained proprietary assets at one or more co-lo facilities. Server virtualization was in early stages of acceptance. Cloud was just over the horizon.
At that point, the three key tenets of IT efficiency being applied to enterprise owned and operated data centers were consolidate, optimize, automate. Try best to simplify, streamline and de-clutter aging IT infrastructures.
Consolidate applied to everything – data centers, storage, networks, applications and especially servers through virtualization.
Optimization brought a payback once consolidation had taken place through tiering of data and applications by importance and resource access. TCO was being reduced by 25% or more, and business continuity was better enabled with more virtualization.
Automation focused on improved forms of deduplication and business process software.
If we were still looking only for the best ways to optimize these traditional enterprise data centers, all of that advice would still make sense. But now there are newer and better options. Most of it centered on virtualization.
All sorts of business is going virtual, and not just IT in the cloud. Think of the obvious examples. Uber, a global ride service, owns no cars. Facebook, a global content service produces no content. Airbnb, a far-flung accommodation provider, owns not a single bed.
The Virtual Data Center
Data virtualization provides the ability to deliver data and applications instantly. Cloud makes it feasible for companies to have all the capabilities of giant data centers without owning a single piece of hardware. The business owns and controls the core value of the data. Data can still be accessed, shared, analyzed, combined, protected and archived. The lifecycle of that data is easily managed through a cloud provider access point that utilizes data anywhere without capital investments.
Virtualized data in cloud deployments can be made both user accessible and secure against any malicious intent. It’s a balancing act substantially simplified by a reduction in excess copies. Virtualization creates fewer copies and presents a smaller attack surface, a tighter span of protection, greater control. The flexibility factor of virtualized data in cloud environments also overcomes the “just-in-case” duplication of more traditional IT structures. It eliminates the excess physical infrastructure created for scenarios like backup, disaster recovery, and business continuity. The resulting flexibility and speed meet business objectives at higher performance levels and lower costs – including substantial energy reductions.
Nevertheless, the pure energy and sustainability impacts of cloud remain to be seen. Cloud has progressed from intense virtualization and private clouds to public clouds, and unified infrastructure and software-defined storage and software-defined data centers and everything as a service. Traditional v. cloud energy consumption is evolving. Questions remain around the energy and sustainability impacts of huge data centers like those that Google, Amazon, and others are building along the Columbia River. Where is the responsibility for the energy and environmental impacts that come from massive concentrations of IT infrastructure? Is it just a set of problems being shifted from traditional data centers to theirs?
These are questions that ultimately tie back to wise planning by all of us as a foundation for how we use technology, or how we let it use us.