X
Forgotten Password?
Retry
Login
Sorry! You can't edit your cart on this page.
Sorry! This item could not be added to your cart as it is no longer available from Comms Express.
Please check the quantity you are adding and try again.
The following item has been added to your cart.
Product Code:
Options:
Qty:
Unit Price:£
View Cart
Checkout
Your cart is empty.
Subtotal:
£0
Learn more about how
to collect Data Points
for free gifts.
Remove all items from cart
Are you sure? Yes No
Request A Quote
View Cart
Apply for a Credit Account
Checkout
Cookie Policy×

Hi there! Our website may store cookies on your computer in order to give you the best experience, such as remembering the items in your cart so you can continue shopping where you left off.

By continuing to use our site, you give consent for cookies to be used.

Browse Categories

Google's AI brings down data centre cooling bill

Google’s total energy use could be slashed by 15 per cent after applying artificial intelligence that manages cooling patterns at its data centres more efficiently than humans.

With billions of web searches to handle every day in addition to powering hugely demanding services like Gmail and YouTube, the internet giant’s servers require a lot of energy to stay cool, so much that they are believed to contribute around two per cent of global greenhouse gas emissions.

However, Google’s DeepMind AI division has now come up with a complicated machine-learning algorithm to accurately anticipate when the servers will be put under increased pressure.

Mustafa Suleyman, DeepMind’s co-founder, said that efficiently handling the servers’ cooling was so complicated and multi-layered to the point that an algorithm was better suited than a human.

“It’s about tweaking all of the knobs simultaneously,” Mr Suleyman explained.

“It’s one of those perfect examples of a setting where humans have a really good intuition they’ve developed over time but the machine-learning algorithm has so much more data that describes real-world conditions.”

The AI has helped Google reduce its energy use for cooling by 40 per cent and total energy use by 15 per cent.

In the past ten years, Google has built its own super efficient servers, employed more efficient cooling techniques, and invested in green energy sources with the eventual goal of being powered entirely by renewable energy.

Google hasn’t revealed exactly how much energy its data centres use but does admit that overall, it is responsible for 0.01 per cent of global electricity use, with its servers accounting for most of that.

Mr Suleyman added that there is still plenty that they can do and that the same algorithm-based approach can be applied in other departments such as power generation facilities and energy networks.

“I really think this is just the beginning. There are lots more opportunities to find efficiencies in data centre infrastructure,” he concluded.

Image: 4X-image/iStockADNFCR-1186-ID-801822447-ADNFCR

Google's AI brings down data centre cooling bill - Article Discussion

Terms & Conditions - This tool is open to everyone. You are responsible for the contents of your comments and any consequences that may arise as a result of them. Comments that contain any of the following will be deleted: terms considered explicit and inappropriate; defamatory, obscene, unlawful, harmful, or offensive.

Join Our Mailing List
 
Social Links
  • Apply for Credit