IT Sustainability Think Tank: Why 2024 is the year for IT managers to revamp their green IT plans

Capture and analyse IT equipment inventory and operational data

The public perception of datacentres darkened in 2023, with growing concerns that these facilities are inefficient energy and resource hogs that will eventually overwhelm available resources. While there is no question that datacentre operations are expanding rapidly, there is very little reliable data on the actual energy consumption and efficiency of IT equipment operations. IT managers need to begin the work necessary to generate and report a work delivered per unit of energy consumed metric.

IT managers are tracking the critical IT equipment inventory, power demand, and capacity utilisation data to monitor the reliability and resiliency of their operations, but only one-third are collecting and aggregating sufficiently granular data to enable the calculation of a work per energy performance metric.

There is a broad ecosystem of software tools, datacentre infrastructure management (DCIM), IT infrastructure management (ITIM), IT operations management (ITOM), and data aggregation software available to craft a comprehensive performance management system. However, deployment will require significant staff resources and one to two years of time.

IT managers need to initiate projects and set a reasonable schedule to implement a data management system capable of calculating reliable work per energy data. This need is particularly urgent in the EU but will become necessary in other jurisdictions over the next five years.

Most importantly, reporting a work per energy metric and energy consumption data will enable managers to communicate the operational improvements and their IT infrastructure's real impact.

Increase IT equipment utilisation and workload consolidation 

The work capacity of much of the IT infrastructure is wasted either by design or laziness. The Uptime Institute 2023 IT and Power efficiency survey found that only 46% of responding IT operators tracked the capacity utilisation of their IT equipment and set utilisation objectives.

Conversations with enterprise operators reveal that many do not consolidate applications when they refresh their IT equipment and are hesitant to deploy power-aware workload management tools out of concern for impacting the reliability of their IT operations.

IT managers must guide their organisations to intelligently utilise their available equipment capacity, using software tools to measure, manage, and maximise utilisation within reliability constraints.

All organizations should set equipment utilisation goals and build multi-year efficiency project plans to improve IT infrastructure energy performance. Data available from IT equipment manufacturers indicate that workloads on two to eight old servers can be migrated to one new server when deploying n+1 (e.g., Intel or AMD CPU generation 3 to 4) or n+2 (e.g., Intel or AMD CPU generation 2 to 4) technology. Similar improvements can be achieved in storage and network equipment.

Consolidating CPU workload and storage and network operations delivers a defined workload on three-quarters to one-half of the equipment. Consolidation significantly reduces capital and operating costs and space, energy, and water use while increasing the work delivered per unit of energy consumed.

Deploy power management functions on servers

IT equipment, particularly servers, are designed with power management capabilities that match equipment power and workload demand, reducing overall energy consumption by up to 10%. These functions are incompatible with all workload types as they increase response time as the CPU switches from one operating state to another or moves into lower power modes when workload is absent.

The response lags may not be appropriate for high-performance compute operations but within the reliability requirements of web serving or enterprise/office applications.

The Uptime Institute 2023 IT and Power efficiency survey found that 54% of IT managers surveyed have an internal policy to enable power management functions, with over half enabling these functions on more than half of their server fleet.

These managers have successfully set internal processes and criteria to govern the enablement of power management functions. In many cases, workloads can be tested for power management compatibility during server testing to verify that performance requirements are met.

IT managers should establish criteria for the controlled deployment of these functions. In the first year, a few projects should be executed for applications/systems where power management is clearly compatible with reliability and response time requirements.

The initial project work enables the organisation to develop experience with power management functions. As experience is gained, more complicated projects can be undertaken, broadening the deployment to further improve energy performance while maintaining workload performance and reliability commitments.

Control IT operations in colocation and cloud datacentres

The Uptime Institute 2022 Environmental Sustainability and Climate Change survey found that only one-quarter to one-third of IT operators tracked operations' energy use and GHG emissions in colocation and cloud datacentres.

The survey results suggest that the energy performance of these operations is not closely tracked and that goals are not established to improve energy performance. Conversations with IT operations support the survey findings, with many IT operators and service providers disavowing responsibility for IT performance. While this is particularly true for colocation operations, it is also true for bare metal and infrastructure as a service (IaaS) cloud operations, where IT operators control energy performance through their BIOS, operating system, and application settings.

IT operators should exercise control of these operations and set goals to improve energy performance. Again, many software tools are available to actively manage these operations to optimise energy performance. Management and optimisation of the utilisation of IT assets in colocation and cloud datacentres can yield operational improvements and cost reductions equivalent to those achieved in owned datacentres.

Increase the consumption of carbon-free energy 

IT managers should establish plans to reduce the carbon intensity of their energy consumption in parallel with their efforts to improve their energy performance.

Typically, these plans are executed by the internal energy management team or the colocation and cloud service providers. They are more likely to be successful when the IT operations team participates in developing the carbon-free energy procurement strategy and supports and tracks its implementation.

Like strategies to improve energy performance, reducing the carbon intensity of the energy supply requires incremental improvements and an extended implementation timeline. Unlike energy efficiency projects, procurement of carbon-free energy will likely have a cost premium of three percent or more. However, having and making progress on a plan to reduce operational GHG emissions enhances an enterprise’s sustainability credentials and facilitates or enables revenue from the increasing number of customers setting their own sustainability objectives.

In conclusion, an effective sustainability strategy delivers business and environmental improvements while maintaining or enhancing reliability, resiliency, and operational performance. With the increased public and regulatory attention to the environmental performance of datacentres, IT managers need to integrate their sustainability strategy into their business plans and make the necessary investments to improve the work delivered per unit of energy consumed metric.

IT managers should start or reinvigorate their sustainability strategies in 2024, working within their organisation and with internal colleagues and critical external suppliers to improve performance through energy efficiency improvements in the IT equipment infrastructure and reductions in the carbon intensity of the energy supply. Projects need to address IT operations in owned, colocation, and cloud facilities.