The ability of IT leaders to leverage powerful public cloud services for digitally empowered business projects and to advance their organizations in the age of artificial intelligence (AI) often results in IT sustainability taking a backseat.
Recent data on cloud expenditure indicates that a rising portion of IT budgets is being allocated to public cloud services. Analysts from Canalys reported in August that global spending on cloud infrastructure services surged by 19% year-on-year in the second quarter of 2024, reaching $78 billion. Despite a rebound in enterprise IT budgets, a large segment is now directed toward AI-related developments.
Shane Herath, chair of the Eco-Friendly Web Alliance, cautions against the environmental implications of tech-driven ventures, particularly as AI technologies demand substantial resources. He states, “The ICT sector accounts for approximately 3% to 4% of global emissions, and data centers consume significant amounts of water for cooling.” He emphasizes that considerable energy consumption leads to notable carbon emissions, and the construction of new data centers to support the increasing demand for AI only aggravates these issues, contributing to water scarcity and depleting finite earth minerals. Additionally, the resulting electronic waste adds to the environmental strain.
Data Center Impact of Digitization and AI
Herath and other industry experts advocate for IT leaders to integrate sustainability into their digital innovation strategies, especially as these initiatives require energy-intensive AI servers and the necessary cooling infrastructure. For over 25 years, driven by the efficiencies of Moore’s Law, IT leaders have been encouraged to manage ever more intricate data center applications and store increasing amounts of data, often neglecting the environmental consequences.
“In this new era of sustainability awareness, we are now obligated to take action based on what has transpired from the turn of the millennium to now,” remarks Mark Molyneux, EMEA chief technology officer at data security and management firm Cohesity. He highlights that corporate IT has benefitted from an age of inexpensive computing and storage, coinciding with the AI renaissance. The emergence of generative AI (GenAI) and large language models (LLMs) is further driving the adoption of hybrid clouds and a broader distributed computing landscape, leading to unprecedented data volume increases.
Measuring IT Consumption
Historically, the expenses associated with IT infrastructure were closely tied to new business demands. Implementing the necessary software and hardware to support IT-enabled business operations usually meant acquiring additional server and storage hardware for new applications. Virtualization led to server consolidation, but in the cloud era, accurately allocating costs has become increasingly complex for IT leaders. FinOps seeks to streamline these cost calculations, with some industry professionals suggesting it should also provide sustainability metrics.
Steve McDowell, chief analyst at Nand Research, believes that energy efficiency metrics are becoming vital for IT infrastructure teams aiming to optimize resources and achieve sustainability objectives. “Active monitoring of power metrics is an exciting new tool for cloud users striving to meet their environmental goals,” he notes.
These metrics can assist data center administrators in adjusting power and cooling levels to align with the business’s environmental goals. However, James Sturrock, director of systems engineering at Nutanix, points out that many enterprise cloud environments are provisioned to operate at fixed grades and levels, lacking the nuanced engineering required for optimal efficiency.
Sturrock poses a critical question: “If cloud application X relies on database Y on service Z and experiences variable load surges and dips that the engineering team may not be aware of, wouldn’t it be prudent to monitor these workflow fluctuations to better align energy-efficiency metrics with IT expenditure plans?”
Nutanix is developing features within its cloud platform to enhance visibility into the power consumption of live production environments. This level of insight can strengthen organizations’ sustainability planning, based on near real-time consumption measurements from active hardware. “Cloud engineers, developers, and operations team members can access straightforward dashboards to visualize power metrics and energy use across their cloud environments,” he explains.
By simplifying the assessment of power consumption, IT departments can obtain precise insights into energy usage rather than relying on estimations or generalized consumption values. Previous efforts to incorporate sustainability metrics into FinOps often depended on spend proxies, leading to inaccuracies. Anodot, a cloud cost management platform, exemplifies a more rigorous approach, having partnered with Greenpixie, a provider of cloud sustainability data, to integrate cost and carbon emissions data into its FinOps tool. Greenpixie’s ISO-14064-verified methodology ensures reliable cloud emissions measurements.
Anodot anticipates that the combined capabilities with Greenpixie will enable its users to access hourly, granular emissions data, providing comprehensive insights into the carbon and water consumption impacts of their cloud strategies. The company emphasizes that carbon awareness can drive engineers to reduce cloud waste even more effectively than financial savings. However, the tools currently available in the FinOps market do not offer the same level of visibility for these two important motivators.
David Drai, CEO and co-founder of Anodot, predicts a growing demand from governments, stakeholders, and customers for accountability in GreenOps, compelling cloud users to adopt more sustainable practices. “Most organizations will need to measure, disclose, and progressively reduce the carbon footprint of their IT operations,” he asserts, foreseeing that the intersection of sustainability and cloud optimization will become standard practice within IT teams and their tools over the next five years.
The Role of IT Consolidation
Metrics can also help IT leaders streamline their infrastructure, preventing the inefficiencies and costs linked to underutilized hardware. Cohesity’s Molyneux advocates for the consolidation of data onto a unified platform, thus minimizing the prevalence of data silos. This can be achieved by focusing on indexing and classifying data according to its relevance and value to the organization, adhering to a predefined record strategy.
By centralizing only necessary data, organizations can better manage resources, while employing techniques like data deduplication and compression. Molyneux asserts that this approach can lead to a nearly 96% reduction in data rates, optimizing storage resources and saving costs while enhancing operational efficiency from the outset.
Sturrock of Nutanix notes that when organizations begin visualizing real-time power consumption metrics alongside historical data to support sustainability goals, IT professionals can achieve an ideal balance between performance and the efficient delivery of applications, data, and computing resources. Technologies such as virtualization, containers, and hyper-converged infrastructure enable workload consolidation on fewer physical devices, thus decreasing energy use and carbon emissions compared to traditional infrastructures.
IT Sustainability and Cost-Efficiency
In addition to consolidation, analyst firm Gartner recommends that IT leaders strive for a 80% storage utilization and 65% server utilization within their data centers. Implementing these practices can yield cost savings of up to 60% and reduce energy demands.
The most impactful action IT leaders can take for both the environment and their budgets is to defer new equipment purchases and enhance the management, optimization, or redeployment of existing assets, according to Gartner’s “Unlock the Business Benefits of Sustainable IT Infrastructure” report.
Autumn Stanish, a director analyst at Gartner, emphasizes, “By extending the lifespan of servers and network devices, organizations can delay new equipment acquisitions, potentially saving up to 40% on related costs. This not only lessens e-waste and the environmental impact of production and transportation but also positively affects the bottom line.”
Finally, Gartner advises IT leaders to set lifecycle standards at the maximum duration that coincides with support from data center equipment manufacturers and to only reduce lifecycles as necessary or when performance requires replacement.