Startup Technology Powers Down Idle Servers
Over a two-month period late last year, Vancouver-based TSO Logic evaluated power consumption of an animation studio’s servers and found that 56 percent of the time the studio was powering idle servers that were not required for their current workload.
TSO Logic performed the evaluation for Arc Productions, a Canadian digital media studio with an in-house server farm of over 600 servers.
TSO Logic’s proprietary technology looks at server activity on an application level, monitoring consumption patterns and finding opportunities to automatically conserve. TSO Logic calls its software “application-aware power management.”
“By correlating power draw, incoming workload (in this case render requests) and application performance, we are able to show the power cost on a per job basis as well as identify wasted power,” said Aaron Rallo, CEO of TSO Logic.
The company’s twin suite of energy efficiency products, called TSO Metrics and TSO Power Control, allows data centers to power down servers automatically when demand is low. TSO Logic estimates it can save ARC Productions more than 800,000 kWh per year for a savings of $88,000 per year.
Most data centers experience variable load, which means that the demand on servers varies from hour to hour. However, to stay prepared for periodic surges-called peak demand or peak load-most data centers keep running at full capacity all of the time. This means that much of the energy consumed is wasted on idle servers. On average, data centers use only 6 percent to 12 percent of the electricity powering their servers, according to consulting firm McKinsey & Company.
- 2013 Insider Knowledge
- Guide to Energy, Carbon and Environmental Software
- Environmental Leader Technology Reviews
- The Logistics, Carbon, and Business Data Book: Fall 2012 Sustainability Trends
- 10 Ways to Slash Energy Cost & Reduce Budget Uncertainty
- EL Insights: Green Plastics
- ISO 50001: Frequently Asked Questions
- Best Practice Guide to Maximize ROI of Energy Management Systems
- REC Databook
- The Business Case for Corporate Sustainability Tools - January 2013