Enablers of a sustainable digital strategyBy Vinod Bijlani
In Singapore, climate change concerns have already impacted technology builds.
Over the past 18 months, industries have had to accelerate their digital transformation journeys to remain competitive. From retail to healthcare and even financial services; no sector has escaped the reality that digitisation is a necessity for survival. Meanwhile, the international calls for urgent action on climate change have made shrinking carbon footprints an additional business imperative. As a result, executives are under tremendous pressure to adapt and transform existing processes to drive customer engagement, employee productivity, and business resilience – all while maximising the sustainability of their IT.
In Singapore, climate change concerns have already impacted technology builds. Experts estimate that data centres consume approximately 7% of the nation’s total energy, prompting the government to put a ‘temporary pause’ on new data centres in 2021. It is estimated that data centres are likely to go from using 1% of global energy in 2021 to increase about 15-fold by 2030, equivalent to 8% of projected global demand, making going green an urgent business for companies worldwide.
So, how should organisations approach miniminsing their IT carbon footprint? The answer is to take a holistic view of the full lifecycles of the solutions that make up their digital transformation projects. This means overlaying a sustainability lens on all they do – from infrastructure, to development and deployment, monitoring and decommissioning.
Underutilisation is not only a business problem but also an environmental one. Often, businesses build in excess capacity into their data centre and cloud resources to ensure optimal performance at any moment. This results in inflated costs and unwarranted energy consumption. While underutilisation estimates vary, according to Computer Economics, nearly 80% of production UNIX servers are utilised at less than 20% capacity, and over 90% of Windows servers are used at less than 20% capacity, while still drawing 30% to 60% of their maximum power.
Thus, when selecting infrastructure providers – whether for on-premises or public cloud – organisations should choose one focused on reducing power usage and carbon emissions. Furthermore, using software defined infrastructure (SDI) and virtualisation techniques helps cut carbon footprint. SDI has emerged as a promising approach to address the extensive demands on maximising the value potential of infrastructure deployments.
Establishing development and deployment processes
Businesses should also examine their software development and deployment processes for sustainability gains, because selecting the proper framework and algorithm can reduce energy consumption. For example, tools like AWS Codeguru help to profile applications and frameworks and provide automatic recommendations on ways to improve. These could include reducing central processing unit (CPU) utilisation and application latency to lower infrastructure costs and enable more energy-efficient applications.
On the deployment front, using microservices-based frameworks are helping with overall infrastructure utilisation. Running applications on serverless technologies can also be highly energy-efficient, as the applications only use the compute infrastructure when they are running.
Another critical decision for organisations is the runtime frameworks for applications. For artificial intelligence (AI) and machine learning (ML) apps, utilising graphics processing units (GPUs) can help reduce the overall carbon footprint. Even though the observed power for GPU is high, applications finish faster, and thus, the total energy consumption is notably less than CPU versions. Additionally, research has shown that GPU achieves an energy-frame reduction ratio of 1.1–3.2× compared to the others for simple kernels, making it a more sustainable alternative.
Monitor and manage applications with AIOps
To avoid excess energy usage, reactions to IT events should be immediate to counter disaster. Traditional approaches to managing IT applications can be heavily manual and result in inefficient energy usage. New techniques like AIOps can enhance the efficiency of managing IT operations and reduce organisations’ digital transformation carbon footprint. AIOps cuts through the noise and identifies, troubleshoots, and resolves common issues within application operations. It brings together data from diverse sources and performs a real-time analysis right at the source. Furthermore, it understands and analyses historical and current data, linking anomalies and observed patterns to relevant events via machine learning. It also initiates appropriate automation-driven action, which can yield uninterrupted and energy-efficient IT operations.
Decommissioning applications and infrastructure
Decommissioning applications is one of the most challenging exercises because of the stored data within the apps. Having the right archival strategy is crucial and will allow IT teams to simplify application decommissioning.
Another essential aspect to consider is infrastructure decommissioning, which needs to be supported by the infrastructure vendor’s circular economy strategy. For example, HPE Financial Services (HPEFS) Asset Upcycling Service is helping customers maximise end-of-use assets, such as Global Airports Group, daa, who converted some of their newer systems into as-a-service using the HPE GreenLake platform, while upcycling obsolete legacy assets, thus benefitting from the residual market value of the old IT while reducing both e-waste and carbon emissions impacts.
The bottom line
With every industry charging ahead with digital transformations, business leaders and environmentalists alike are concerned with the negative impacts of surging energy consumption. As industries everywhere mobilise to embrace the digital economy, smart companies will seize this moment as an opportunity to integrate sustainability into their digital transformation strategies – to mitigate both the societal and financial costs of unchecked energy demands.