FI
Summary
We have developed a method for assessing the energy efficiency of information systems and optimizing systems to reduce energy consumption. During the development of this method, we built an audit process and a template for documentation, and gathered extensive knowledge materials to support future audits and optimizations.
For the pilot phase audits, we selected four companies that were sufficiently diverse from the perspective of developing the audit process. The criteria included the size of the company and product development organization, the technologies and platforms used in the systems, and the amount of web traffic occurring in the systems.
The total estimated annual savings potential for all pilot audits was €120,000 in financial savings, 2 tons of CO2 emissions, and 7 MWh in electricity consumption. The optimization recommendations based on these estimated figures were fully aligned with the current architecture and resources of our clients' systems. The savings potential does not include comprehensive renewal proposals that would yield the most significant savings, as each major renewal requires a more detailed development plan.
The majority of emissions from information systems come from indirect emissions (Scope 3) within the value chain of data centers and service providers. A simple example of such emissions is new hardware acquisition for a data center. Recent technological advances in cloud and virtualization technologies make it significantly easier for developers to reserve the resources they need from data centers, which can result in growing hidden costs and emissions for the company.
In two of the pilot audits, light optimization was performed while monitoring the impact of the optimizations. In one, emissions were reduced by a third of the original value, and costs were reduced by 10%. Prior to the optimization, we estimated a broad savings range, with costs ranging from 30% to 60% and emissions from 50% to 70%, which can be achieved by implementing the remaining optimization recommendations.
On a larger scale, the problem is much more significant, especially when agreed-upon common practices are not in place or monitoring is not conducted. Unused, reserved resources were found in two pilot audits. In three of the pilot audits, it was found that the used resources were very underutilized, except for occasional load spikes—mainly, the systems were active during office hours.
All audited systems showed potential for architectural improvements to significantly increase savings in terms of costs and emissions. The most significant optimization recommendation is to use cloud-native solutions or technologies that enable systems to respond adaptively to different load situations—scaling down reserved resources during low load or shutting them down completely, and increasing resources as needed during high load. Automatic scaling alone is not the most effective way to build sustainable systems; rather, each system component needs to be evaluated individually. It is important to know which functionalities each system component is responsible for and how much it can utilize resources.
Throughout the spring and summer, we will publish practical observations with examples from our audits in our energy efficiency blog series, which can be found via the following link:: https://www.atoz.fi/en/blogi/tag/energyefficiency.