Featured Research

from universities, journals, and other organizations

Putting more cores to work in server farms

Date:
November 26, 2012
Source:
Ecole Polytechnique Fédérale de Lausanne
Summary:
Scientists have found that reorganizing the inner architecture of the processors used in massive data processing centers can yield significant energy savings. Their work is part of the EcoCloud program.

EPFL scientists have found that reorganizing the inner architecture of the processors used in massive data processing centers can yield significant energy savings. This is a research from EcoCloud research center, which was founded to pioneer technologies to make cloud computing scalable, cost-effective and sustainable.

Streaming data, social networks, online games and services, databases -- the number of interactions we have with the Internet is continually increasing. Every time we click on a link, we trigger an avalanche of computer operations that are then carried out in huge server farms. It's estimated that these massive installations are responsible for 2% of total world electricity consumption. EPFL EcoCloud Scientists are proposing a novel solution to help rein in this runaway consumption. By integrating the same kind of processor cores that are used in smartphones, the amount of energy needed can be reduced by a factor of four. Their study is part of EcoCloud, a research center dedicated to cloud computing. The results were recently published in an IEEE Micro article.

The giants of the digital world -- such as Facebook, Google, Microsoft -- all depend on vast, powerful farms with tens of thousands of servers to manage their data processing. To help keep costs down and to improve energy efficiency, chips have been improved and packed as tightly as possible into the processors. But this approach has reached its limits.

EcoCloud's solution, titled "scale-out processors," is based on a different approach. They propose a reorganization and redesign of the processors used in the servers. Instead of the current design, which is based on a few, very powerful processor cores, they recommend using a greater number of less powerful cores. Each processor could thus respond to a larger number of requests.

Over-powerful

"The vast majority of Internet requests don't involve complicated analysis, but are generally just retrieval from memory," explains Boris Grot, from Parallel Systems Architecture Laboratory (PARSA). "But current servers are designed for carrying out a whole range of tasks, from complex scientific calculations to gaming. They're actually way too powerful for most basic demands. As a result, they're not being used in an optimal manner."

The researchers have combined the advantages of new-generation small processor cores developed for smartphone-type devices; their architecture is simple but their processing ability is very efficient. Concentrated in large numbers in a large chip, they would provide a better solution to the way servers are currently used. After having studied and compared several designs, EcoCloud scientists concluded that this arrangement maximizes space in the processors and significantly improves their performance.


Story Source:

The above story is based on materials provided by Ecole Polytechnique Fédérale de Lausanne. The original article was written by Sarah Perrin. Note: Materials may be edited for content and length.


Journal Reference:

  1. Boris Grot, Damien Hardy, Pejman Lotfi-Kamran, Babak Falsafi, Chrysostomos Nicopoulos, Yiannakis Sazeides. Optimizing Data-Center TCO with Scale-Out Processors. IEEE Micro, 2012; 32 (5): 52 DOI: 10.1109/MM.2012.71

Cite This Page:

Ecole Polytechnique Fédérale de Lausanne. "Putting more cores to work in server farms." ScienceDaily. ScienceDaily, 26 November 2012. <www.sciencedaily.com/releases/2012/11/121126131211.htm>.
Ecole Polytechnique Fédérale de Lausanne. (2012, November 26). Putting more cores to work in server farms. ScienceDaily. Retrieved April 18, 2014 from www.sciencedaily.com/releases/2012/11/121126131211.htm
Ecole Polytechnique Fédérale de Lausanne. "Putting more cores to work in server farms." ScienceDaily. www.sciencedaily.com/releases/2012/11/121126131211.htm (accessed April 18, 2014).

Share This



More Computers & Math News

Friday, April 18, 2014

Featured Research

from universities, journals, and other organizations


Featured Videos

from AP, Reuters, AFP, and other news services

Facebook Announces Location-Sharing Feature 'Nearby Friends'

Facebook Announces Location-Sharing Feature 'Nearby Friends'

Newsy (Apr. 18, 2014) — Facebook's pending Nearby Friends feature will give users the option to share their nonspecific or specific locations with certain friends. Video provided by Newsy
Powered by NewsLook.com
Michaels Hack Compromises About 3 Million Credit Cards

Michaels Hack Compromises About 3 Million Credit Cards

Newsy (Apr. 18, 2014) — Michaels is now confirming that an eight-month security breach compromised about 3 million customers' credit and debit card data. Video provided by Newsy
Powered by NewsLook.com
Twitter Introduces Facebook-Style App Install Ads

Twitter Introduces Facebook-Style App Install Ads

Newsy (Apr. 17, 2014) — Twitter hopes to make money on app install ads, which has proven to be a successful strategy for Facebook. Video provided by Newsy
Powered by NewsLook.com
Heartbleed Hack Leads To Arrest

Heartbleed Hack Leads To Arrest

Newsy (Apr. 17, 2014) — A 19-year-old computer science student has been arrested in relation to a data breach of 900 social insurance numbers from Canada's revenue agency. Video provided by Newsy
Powered by NewsLook.com

Search ScienceDaily

Number of stories in archives: 140,361

Find with keyword(s):
 
Enter a keyword or phrase to search ScienceDaily for related topics and research stories.

Save/Print:
Share:  

Breaking News:
from the past week

In Other News

... from NewsDaily.com

Science News

Health News

Environment News

Technology News



Save/Print:
Share:  

Free Subscriptions


Get the latest science news with ScienceDaily's free email newsletters, updated daily and weekly. Or view hourly updated newsfeeds in your RSS reader:

Get Social & Mobile


Keep up to date with the latest news from ScienceDaily via social networks and mobile apps:

Have Feedback?


Tell us what you think of ScienceDaily -- we welcome both positive and negative comments. Have any problems using the site? Questions?
Mobile iPhone Android Web
Follow Facebook Twitter Google+
Subscribe RSS Feeds Email Newsletters
Latest Headlines Health & Medicine Mind & Brain Space & Time Matter & Energy Computers & Math Plants & Animals Earth & Climate Fossils & Ruins