Database optimization and the importance of reducing global digital pollution

I’ve been thinking back to Postgres Ibiza (PGIBZ) that took place a few weeks ago… And not just because I miss the warmth and sunshine of Ibiza as winter as arrived in Cognac! No, I was remembering how the island was still bearing some signs of the recent floods, even though the islanders were actively working on erasing and repairing the impacts it had caused. The latest news from the Caribbeans and from Asia made me realize that Ibiza had been, relatively speaking, quite lucky.

That particular atmospheric event followed major heatwaves and broken records across the U.S., Europe, and Asia in the past few years. Just this last week at the time of this article, there has been historic flooding with profound damage in Washington state (U.S.)… more than nine million acre-feet of water dumped on the state over ten days… with the torrential downpour predicted to continue throughout the West Coast the rest of the year.

I don’t know if the relevancy of Ibiza’s floods in particular played a part in the way my short talk at the conference was received, but I am still grateful the topic at hand was not only heard, but listened to - even thought about and discussed. I must admit I was not expecting that.

I am part of those who believe that global environmental challenges are about to completely transform our societies on every level as resources get more limited and redistributed. I am the one asking around if anyone else wonders “what good will my comfortable salary do me if potatoes can’t grow anymore?”. Maslow clearly illustrated that it is less necessary to be rich than fed. Yet, as we all know, the environment is still not priority one, even after all this time and all the lessons we’ve experienced.

So, for me it made sense to discuss the topic at a PostgreSQL conference, because it should simply be discussed everywhere. It is a topic that is relevant to each and every one of us, because it involves the Earth itself that we reside on and take for granted. Yet, was I confident about finding a receptive audience? Nope!

I felt slightly uncomfortable – I usually do as a non-tech person at a technical event… – for taking the audience into a space they were not necessarily there for. I was hoping I would not bore, or bother, or even annoy. But after only the first few slides I saw that people were taking pictures of the data I was sharing, some were nodding, some were seeing those figures for the first time and raised an eyebrow. And when I linked it all with what we do daily and the impact of the digital pollution we produce, instead of facing disinterest or disapproval I got “why not!” looks in the room.

So what digital pollution do we produce?

According to Greenpeace, digital pollution refers to all forms of pollution caused by the IT sector: greenhouse gas emissions, chemical contamination, erosion of biodiversity, and production of electronic waste.

Yes, there are some innovative solutions being designed to handle some of this digital pollution. For example, some situations are ideal for creating data centers that are able to use the heat they produce in order to heat nearby offices or homes. Projects such as Amazon’s Tallaght project is “the first large-scale district heating network of its kind in Ireland”, helping to “significantly reduce emissions in the area by almost 1,500 tonnes of CO2/year, establishing Tallaght as a leader in local energy action” (source).

However, the growth in the number of users equipped with at least one connected terminal (especially in developing countries), the increase in the ratio of the number of connected terminals per individual (from 2.1 in 2015 to 3.3 in 2020 on average worldwide), the increase in video traffic coupled with the increasing share of HD and UHD quality images and the shift of usage to consumption on demand (streaming, VOD, cloud gaming) have resulted in an explosion of traffic on networks (more than 25% per year) and in data centers (more than 35% per year). This growth is occurring at a rate that surpasses energy efficiency gains in equipment, networks and data centers. These traffic forecasts are also regularly revised upwards.

Although the number of “standard” devices (excluding connected objects) has increased significantly between 2000 and 2015, it stabilises between 2015 and 2025 because the market is saturated.

Although the mass per user decreases slightly, from 63 kg CO2 emmissions to 58 kg between 2010 and 2025, the total mass (user equipment, networks, data centres) is multiplied by more than 2.5 in 15 years: it increases from 128 million tonnes in 2010 to 317 million tonnes in 2025.

This explains the stress on raw materials, especially “conflict minerals” and other rare earths.

The key to a greener world

Performance optimization for databases doesn’t just result in improved costs and speed. When we work on improving a query, we save that little bit of extra energy that would have been used to read/write, that extra disk movement that was always unnecessary. Other essential responsible practices that in a clever way allow you to better harness the power of technology results not just in a more technically efficient result but also - without realizing or thinking about it - doing some good for the planet.

For example, some actions we commonly consider for our own good that have an equally beneficial result for our planet include:

query optimization
index optimization
conducting performance audits to see bottlenecks and identify problems
selecting less energy-intensive technologies
reducing your data footprint (removing redundant, corrupt, or unnecessary data)
optimizing the storage infrastructure and computing power to reduce energy consumption

We need to think about digital sobriety, thoughtful coding practices and considerate energy consumption as being the key to greener coding and integrate that parameter in the way we work every day and in the way we build, manage and develop database solutions in order to minimize our environmental footprints as much as possible. These kinds of actions can generate significant cost savings for our own companies and for our customers, which can then be reinvested in R&D projects to increase energy efficiency. It’s a virtuous circle.

A closing thought to this section: AI in particular drives far higher water usage, emissions, and e-waste (training AI models is one of the most resource-intensive computing tasks on the planet!). In a world where it has become common to integrate AI into every corner of applications and day-to-day coding practices, being deliberate in how such technologies are used to ensure they are truly the best solution for the problem at hand has a huge impact. Is AI really the most appropriate solution for that issue you’ve been thinking about? Or are you considering its usage merely because it’s easier to implement than something more technically involved?

Back to you

And that’s an interesting twist to consider regarding our day-to-day. We sometimes complain about needing to know there’s meaning to what we do; well here’s some food for thought to fuel your complaints!

A brief note from our Marketing team here at Data Bene: our team of experienced PostgreSQL engineers can help with improving your own energy efficiency, database optimizations, and thoughtful coding practices that will result in direct benefits to you: speedier database operations, decreased resource usage, lessened costs, and a greener planet. Curious about what we offer? Get in touch - we’d love to chat.