Share this
Celebrating Earth Day with Infegy’s year-on-year energy efficiency achievements
by Hephzibah Dutt on April 20, 2023
As a society, it took us a little too long to face the adverse environmental effects of digital technologies. In fact, it wasn't until the high energy consumption of cryptocurrency mining hit mainstream conversations (circa 2018) that most IT industry leaders even considered the hidden ecosystems and complex carbon footprint of digital infrastructure!
Today, these conversations – particularly the energy consumption of data centers and digital data processing industries – have finally gone mainstream, and many industry leaders are working retroactively to reduce their environmental impact. Here at Infegy, however, efficiency practices and ecological care were hardwired into the foundational building blocks of the company.
In commemoration of Earth Month, we sat down with Infegy Founder and CEO Justin Graves (pictured below) to learn more about his vision for an environmentally responsible tech company and the sustainability practices he’s been iterating on since he founded the company in 2007.
Taking control of our carbon destiny
HD: You’ve spoken to the company at large about your aspiration for Infegy to be a “True Zero” company in terms of our greenhouse gas emissions. Why is that a critical aspiration for you/us, especially since we’re “just” a social intelligence and data company?
JG: I think climate change is the biggest risk factor ever to face humanity. Data center power consumption is phenomenally high. Server energy is becoming an increasingly big deal! Bitcoin mining really brought this to people’s attention.
The US Department of Energy has shared this: around 2% of the energy in the USA is consumed by data centers – and this is probably underreported because it doesn't even account for … let’s say, businesses like ours. What’s illuminating though, is that 2% energy consumption is about equivalent to the US aviation industry’s 2% impact on CO2! So if we’re all not trying to do something about it, to me, that’s a huge error as human beings.
HD: Yes. So what are one or two things we’re doing at Infegy to address that?
JG: We’re taking control of our carbon destiny!
For Infegy, our efforts are centered around energy usage – electricity usage in particular – from the servers we operate ourselves. This is pretty unique in the data and analytics software space. Unlike many consumer intelligence or social listening providers, we have complete control of the energy efficiency of our servers. We can put them in places that need less cooling, and our team monitors them daily. We took control of the fan systems and the configuration of our racks and even reduced the number of network switches we use.
Again, because we own and operate our own servers, we conducted extensive lab testing to determine the highest safe temperature at which our CPUs can idle. Server fans spin up to jet engine speeds; imagine dozens of those and the energy it takes to power them. So now, we can let our fans run at lower speeds, and that can save 100 W across the room – just by slowing the fans.
Talking about temperatures, we switched our hard drives out as soon as the tech became available. Eventually, we rebuilt our servers using fancy, highly efficient power distribution systems. Nowadays, we can regulate the temperature in the machines by varying fan speeds really effectively in that thermal cycling doesn't affect NAND storage and solid-state devices anywhere near as much as it would hard drives – those required us to keep our rooms cooler. The power distribution alone allows us to save around 5% of our energy consumption, and every little bit helps in this situation. I can share numbers here in a bit.
Speed to insight and the race to idle
HD: I’ve heard you talk about how the speed of Infegy Atlas – your efforts to make sure it processes queries in seconds – has bearing on our energy efficiency practices as a company. Could you tell us more about that?
JG: Yes! This is part of the DNA of Infegy: what we've done, have always done, is try to make our software as fast and efficient as possible.
This, of course, has had a dual benefit: we promise our customers speed to insight, that regardless of the query, Infegy Atlas will process in seconds.
I am an efficiency-obsessed person in the world of software. The fact that our systems are so efficient is extremely helpful when you care about efficiency in the environment and sustainability.
So to that end, if we can say, hey, our software needs 1/10th, 1/100th – or in many cases, legitimately – 1/1000th of the compute resources to do its task versus a comparable alternative, that means the computer itself, the processor in particular in that computer has to increase its energy consumption.
[At this point, JG waxes technical, and the IT and computer science nerds in the room ply him for details regarding data transfers, batches, configurations, clock speeds, and memory tests that he and his team devised to push the computers in the race to idle. 10 minutes later… ]
HD: So, in lay-person terms, you did all this in efforts towards “the race to idle”: you want the computer to finish its work as fast as possible and get back to idle as quickly as it can so that it will use less power.
JG: Yes. So, not only does Infegy Atlas get done processing queries faster than any other social intelligence platform on the market, we’re seeing gains all over the place in terms of energy consumption.
Related: Learn more about how Infegy Atlas can get you consumer intelligence at speed.
Image: Justin Graves and the Operations Team performing systems maintenance.
I'll give you some numbers that are pretty interesting.
We've really been on this crusade – beyond strictly just efficiency in software – since late 2020. We’ve been tracking how much total energy the servers and all their supporting ancillary equipment are using on a daily basis.
Now, in March of 2021, we deployed our new, way more efficient, infrastructure. We were able to reduce the equipment we were running quite a bit and dropped consumption by nearly 24%. And then, if you go to the year after this, our average usage decreased by 14% in 2022 - purely from improvements with our use of that hardware.
But this is where it gets really interesting: between 2022 and 2023, the amount of data we're storing went up, with the amount we're collecting per day increasing by 20%. Furthermore, the amount of users and activity we have on our platform has increased by roughly 40%. And yet our average daily energy usage for February 2023 went down 14% from last year!
So to me, that's a “money where the mouth is” kind of number, right? We can increase the data we have, we can increase the rate at which we're collecting it, we can increase users and activity on the platform … and somehow, simultaneously decrease power consumption.
As I mentioned earlier, it is a crucial goal of the business to get to True Zero – not Offsetting Net Zero – True Zero. There are many things I aspire to for the office and company; we’re currently figuring out ways to make the building more efficient and looking into solar, but while we are based in Kansas City, running the servers off renewables is much more challenging. We have a plan in the works, though … so stay tuned.
In the meantime, going full-tilt at what we can realistically accomplish – the low-hanging fruit, per se – has had tangible rewards that I’m really proud of.
It’s shown us that when we focus on efficiency, the impossible becomes possible.
Share this
- December 2024 (1)
- November 2024 (1)
- October 2024 (2)
- September 2024 (3)
- August 2024 (4)
- July 2024 (2)
- June 2024 (1)
- May 2024 (2)
- April 2024 (2)
- March 2024 (3)
- February 2024 (3)
- January 2024 (2)
- December 2023 (3)
- November 2023 (4)
- October 2023 (3)
- September 2023 (3)
- August 2023 (4)
- July 2023 (4)
- June 2023 (2)
- May 2023 (4)
- April 2023 (4)
- March 2023 (4)
- February 2023 (4)
- January 2023 (1)
- December 2022 (3)
- November 2022 (4)
- October 2022 (3)
- September 2022 (3)
- August 2022 (2)
- July 2022 (1)
- June 2022 (1)
- April 2022 (1)
- March 2022 (1)
- January 2022 (1)
- December 2021 (1)
- November 2021 (1)
- October 2021 (1)
- June 2021 (1)
- May 2021 (1)
- April 2021 (1)
- March 2021 (1)
- February 2021 (1)
- January 2021 (2)
- November 2020 (1)
- October 2020 (2)
- September 2020 (1)
- August 2020 (2)
- July 2020 (2)
- June 2020 (2)
- April 2020 (1)
- March 2020 (2)
- February 2020 (2)
- January 2020 (2)
- December 2019 (2)
- November 2019 (1)
- October 2019 (1)
- September 2019 (2)
- August 2019 (2)
- July 2019 (1)
- June 2019 (1)
- May 2019 (2)
- March 2019 (2)
- February 2019 (2)
- January 2019 (1)