IT Sustainability Think Tank: Closing the sustainability gap takes patience and persistence

Organisational commitment

Successfully implementing an IT system sustainability strategy requires commitment from the CIO and CFO to the operations teams. Critically, it must be a long-term commitment to a defined, resourced plan. Too often, sustainability is an initiative ‘de jour’ to be relentlessly flogged across the organisation for twelve months with urgent meetings and high-priority programs that are allowed to fade away in the face of new, more pressing initiatives. Organisations that have achieved transformational changes in their work delivered per MWh consumed take years to transition, and the job is never complete due to constant innovation in IT equipment and software management systems.

The collection of operational data including equipment counts, capacities, utilisation and power demand, equipment and system level energy and water consumption, and equipment utilisation profiles, must anchor the commitment. This data is used to generate key sustainability metrics such as applications, virtual machines, and containers per MWh or server, average CPU or memory capacity utilisation, percent of available storage TB used, computer room air handler (CRAH) cooling utilisation and average IT space temperature. Each metric should have a corresponding sustainability goal to drive environmental performance improvements.

The metrics and progress towards their associated goals should be reviewed periodically across the organisation, with regular reviews at the operational level and four-month reviews at the executive level. Goals should be updated as targets are achieved, new opportunities for improvement are identified, and new technologies are deployed.

Continual improvement

When viewed in its entirety, a sustainability strategy can appear as an insurmountable obstacle given the resource demands, the potential impacts on resiliency and reliability, and the sheer quantity of projects. The IT operations team must create an overall plan that prioritises projects for implementation within each efficiency category (IT equipment consolidation, IT space temperature increases, etc.) over a period of several years. The timeline of the plan should be constructed to match available skills and resources and develop needed technical expertise and standard operating processes. Consider two examples of incremental execution of IT efficiency projects.

Most IT organisations have opportunities to increase IT equipment utilisation but may not have the resources and skills to support stand-alone server consolidation projects. The team decides to leverage the refresh cycle to drive consolidation and sets an objective of consolidating an average of two existing servers onto each refreshed server. External skills are hired to facilitate the rewrite or replacement of selected software applications to enable consolidation. The IT and procurement teams collaborate to purchase a workload consolidation package to guide the consolidation process. As a consolidation project will likely reduce the datacentre space required by the new servers by 25 to 50%, additional efforts will need to be coordinated with the facilities team to consolidate the IT equipment and rebalance and optimise the cooling delivery.

Project success will be measured by increased average utilisation of the IT equipment, reduced capital expenditures for the refresh, decreased software license and system administration costs, and increased work delivered per MWh consumed. One or two projects should be performed in the first year, with the learning from the project(s) applied to facilitate subsequent projects.

Many legacy datacentres have room to increase the IT space inlet temperatures using a monitoring and control software package. Project implementation relies on the support of three internal organisations. Procurement needs to select suppliers for temperature and pressure sensors, the software package, and an installer/integrator.

Facilities operations will integrate the chosen system into their existing control/alarm systems, operate it, and manage the incremental increase to IT space temperature to the new set point. IT operations must agree to the temperature increase process and monitor the IT infrastructure for signs of stress.

Project success is measured by the decrease in Power Usage Effectiveness (PUE), the increase in average datacentre temperature, the increased utilisation of the CRAH units, and reductions in operating expenses. Here again, learning obtained from executing the initial project should be applied to drive efficiencies and improvements in subsequent projects.

These are just two examples of the many efficiency projects available to improve datacentre environmental performance. As many IT organisations have not begun to build their sustainability strategy, it is critical to get started and embrace these three crucial points.

1. Executing a sustainability strategy is a journey, not a sprint. The required personnel, skills, and funds are finite, so projects must be prioritised over several years. Early projects are selected based on their ability to deliver a high return, build required technical skills and experience, and generate organisational momentum.
2. The entire enterprise – the procurement, IT, facilities groups, and colocation and cloud providers where they are used - must be involved and committed to the execution of the strategy.
3. To succeed, the projects must improve business and environmental performance