Research

Publications

Prospective Students

Events

People

Search

Seminars

Links

Home

2013 -14 Seminars


Green Computing: Energy Efficiency at Multiple Scales

Yuvraj Agarwal
Assistant Professor, School of Computer Science
Carnegie Mellon University

Abstract
Our societal energy consumption is rising at a staggering rate. Part of this problem is due to the rise in energy use, and in turn the carbon footprint, of Information Technology equipment worldwide (devices, servers, networking equipment). Traditional mechanisms to save energy such as duty-cycling by putting computing platforms and their subsystems into power saving sleep modes donít work well in practice due to several application scenarios that require maintaining continuous network presence. I will describe two architectures, Somniloquy and SleepServers, that we have developed that enable IT equipment, such as PCs, to be in a hybrid state of operation where they are able to enter sleep modes while maintaining network connectivity. Additionally, computation can be performed on behalf of the sleeping PCs, allowing for even greater savings. This "sleep-talking" state of operation can save, on average, 70% of the energy consumed by a typical PC. While computing is indeed part of the problem due to its increasing carbon footprint, in the second part of my talk, I will show that computing is also part of the solution, where it can be used to make other systems much more energy efficient. In particular, I will focus on sensing and control solutions that we have designed and deployed within enterprise buildings to make them more energy efficient and sustainable. I will show that by using fine-grained occupancy information gathered either from battery powered wireless sensors, or from smartphones and WiFi Access Points, the energy consumption of the HVAC system within a building can be reduced dramatically, saving up to 40% in a test deployment. I will also describe our smart energy meter that can measure the energy usage of plug-loads within a building as well as provide a mechanism to control these loads based on a number of policies.