Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

Author ORCID Identifier



Open Access Dissertation

Document Type


Degree Name

Doctor of Philosophy (PhD)

Degree Program

Computer Science

Year Degree Awarded


Month Degree Awarded


First Advisor

Ramesh Sitaraman

Second Advisor

Prashant Shenoy

Subject Categories

Digital Communications and Networking | OS and Networks


Internet-scale distributed systems such as content delivery networks (CDNs) operate hundreds of thousands of servers deployed in thousands of data center locations around the globe. Since the energy costs of operating such a large IT infrastructure are a significant fraction of the total operating costs, we argue for redesigning them to incorporate energy optimization as a first-order principle. We focus on CDNs and demonstrate techniques to save energy while meeting client-perceived service level agreements (SLAs) and minimizing impact on hardware reliability. Servers deployed at individual data centers can be switched off at low load to save energy. We show that it is possible to save energy while providing client-perceived availability and limited impact on hardware reliability. We propose an optimal offline algorithm and an online algorithm to extract energy savings and evaluate them on real production workload traces. Our results show that it is possible to reduce the energy consumption of a CDN by 51% while ensuring a high level of availability and incurring an average of one on-off transition per server per day. We propose a novel technique called cluster shutdown that switches off an entire cluster of servers, thus saving on both server and cooling power. We present an algorithm for cluster shutdown that is based on realistic power models for servers and cooling equipment and can be implemented as a part of the global load balancer of a CDN. We argue that cluster shutdown has intrinsic architectural advantages over server shutdown techniques in the CDN context, and show that it outperforms server shutdown in a wide range of operating regimes. To reduce energy costs, we propose a demand-response technique that responds to pricing signals from a smart grid by deferring elastic load. We propose an optimal offline algorithm for demand response and evaluate it on production workloads from a commercial CDN using realistic electricity pricing models. We show that energy cost savings can be achieved with no increase in the bandwidth cost.