Cloud-Native Data Analytics: An Eco-Friendly Solution

Cycle to work schemes; vegan snacks; replacing flights with zoom calls. What is the best way organizations can take action on climate change?

One sector that is often overlooked is IT, specifically with regards to how organizations store and manage their data. Migrating data to the Cloud delivers massive improvements in the areas of efficiency and sustainability. While traditional IT infrastructure relies on on-site servers and is hugely energy inefficient, meaning higher costs and carbon footprint, Cloud-based platforms and solutions offer opportunities for improved and eco-friendly operations.

In this article, we look at different aspects of the data ecosystem, and how Cloud-native Business Intelligence (BI) solutions produce superior insights and even reduce dangerous emissions.

Data Storage

Definition: ‘…the retention of information using technology specifically developed to keep that data and have it as accessible as necessary.’ – Hewlett Packard Enterprise

Optimizing how data is stored reduces energy use and overhead costs: a win-win. According to Gartner, server utilization at the average data center is often below 50%, and even as low as 20%. This is an issue as servers constantly consume electricity. Under-utilization is especially prevalent with on-site data centers. Organizations must purchase spare capacity in anticipation of server usage spikes – the surplus requires energy and, in most cases, that means carbon emissions. On the other hand, running a Cloud-based BI tool is two-to-four times more efficient than traditional data centers.

Decommissioning one server can save US$500 in energy, US$500 in operating system licenses, and US$1,500 in hardware maintenance costs each year, according to estimates by Gartner. Higher-efficiency cooling techniques can further reduce energy consumption; solutions can be as simple as implementing a hot aisle/cold aisle configuration for the servers.

Colocation, the practice of renting servers at a third-party provider’s data center, can further optimize data storage. First, underutilized centers could be closed, with the storage requirements taken on by other centers with spare capacity. Second, demand can be directed towards green data centers, encouraging the transition towards renewable energy. These are both arguments for migrating data to the Cloud. Astrato can then deliver BI directly from the Cloud, meaning powerful, up to date insights: playing the data where it lives.

Data Management

Definition: ‘…the practice of collecting, keeping, and using data securely, efficiently, and cost-effectively.’ – Oracle

Data storage receives most of the attention in discussions around sustainability, but data protection and management is also costly, amounting to 70% of data management spend, according to Sustainability Magazine. Minimizing the amount of ‘cold’ data – data that is rarely accessed – being managed on expensive, high-performance storage and backup systems goes some way to addressing this issue. Better still, avoid the generation of unnecessary data with a rigorous data retention and deletion policy. There is huge scope for this kind of initiative; in most organizations, at least 60% of data has not been touched in months, or even years. Cloud-based data management, where the data is in one place, makes it easier to differentiate between useful and surplus data.

Data Architecture

Definition: ‘…the policies, rules, standards, and models that govern data collection and how data is stored, managed, processed, and used within the organization.’ – Snowflake

It’s safe to say that Artificial Intelligence (AI) and Machine Learning (ML) are here to stay. Along with the revolutionary technological advances these systems promise, comes an environmental burden. But one that can be mitigated and even avoided. The carbon footprint of training and using such technologies can be reduced by running the algorithms in different parts of the world, and at specific times of day, to optimize energy efficiency. To give a specific example: training a ML model in Estonia may emit up to 61 times more carbon than training the same model in Sweden. Similarly, training models during hours where the carbon intensity of the energy is lower can massively reduce emissions; in the cases of Denmark and Britain, by 75% and 50% respectively.

And write good code! Of course, this is easier said than done, but making algorithms as efficient as possible minimizes computing power and therefore energy use.

How Astrato Can Support Businesses to Reduce their Carbon Footprint

Collaboration is essential to meaningful action on climate change. Gartner recommends establishing a baseline footprint by aggregating and tracking data on computing-related emissions. Astrato facilitates both these initiatives, offering a Cloud-native, interactive solution, providing accessible, actionable insights that highlight inefficiencies and monitor progress.

After measuring environmental impact in real time, establish specific and measurable IT sustainability goals and measure progress, with an accountability framework and formal policies for achieving these goals.

Case Study

The Kellogg Company (of Cornflake fame) is a valuable case study. In 2005, the food manufacturer employed real-time data monitoring, alongside a data management platform which tracked energy use. The key finding was a massive inefficiency in how it was controlling the temperature and ventilation of its headquarters (some of the systems were working against each other). In fixing this issue across its sites, the company prevented close to 400,000 metric tons of carbon from entering the atmosphere over seven years, not to mention US$3.3 million annual savings.

The first step in achieving these massive environmental and cost-saving benefits was monitoring energy-use data. An Astrato dashboard can deliver this, along with powerful BI insights, ensuring organizations measure what matters and take action accordingly.

Astrato also makes data easier to manage. Via the intuitive interface and AI Insights tool, business users can query data independently. This makes environmental sense in two ways. The first is direct: a business user analyzing the data means there is no back-and-forth with an analytics team, reducing computational energy use. The second is that all members of an organization can view and understand progress on the group’s sustainability goals.

Organic, bottom-up action is always more powerful than bureaucratic sustainability policy. Astrato encourages this kind of collaborative ethos; individuals can see the results of their action and progress towards their shared goal.

Ready to use Astrato to achieve your sustainability goals? Book a demo to find out more, or check out how Astrato can deliver business intelligence across a range of use cases through our example workbooks.