This project addresses the reduction of energy consumption in optical and wireless access networks via a series of approaches such as capacity-adaptive network design, and energy-efficient resource allocation and traffic scheduling. The capacity-adaptive access network stays in the 'high-capacity high-power' mode when the network is heavily loaded with bandwidth-hungry applications such as video streaming, and switches into the 'low-capacity low-power' mode when the network is lightly loaded with less bandwidth demanding applications such as voice and future smart grid traffic. The project generalizes the approaches currently being used in data centers to communication links, and focuses on formalizing the problem and obtaining optimal or near-optimal solutions. It will not only improve energy efficiency of current access networks, but will also provide insights and guidelines on how to upgrade the capacity of access networks efficiently.

The concept of designing capacity-adaptive access networks introduces new requirements of 'capacity adaptive' on the design of access networks as opposed to simply provisioning for peak utilization. This project's energy-aware network control, resource allocation, and traffic scheduling schemes consider an additional key performance metric, energy consumption, in addition to network quality of service, while conventional schemes merely consider network performance such as delay and loss. Research results, via theoretical analysis, simulations, and testbed experiments, will not only demonstrate improved energy efficiency of access networks, but will also be potentially tailored for core networks, datacenter networks, and other wireless systems.

Broader Impact: Reducing the energy consumption of access equipment will not only have positive environment impacts during normal operation but it will provide public benefit during emergencies. Remote access equipment is currently powered by a combination of public power, battery backup, and varying degrees of local generation. Techniques such as those being developed by this project that extend operation when on emergency backup help protect the public health and safety. The project will integrate research and education through participation of both undergraduate and graduate students in the project and by incorporation of research outcomes into undergraduate and graduate course work. The PI will utilize the New Jersey Educational Opportunity Program (EOP), which provides educational opportunities to under-represented populations, as part of the process for recruiting students into the research program.

Agency
National Science Foundation (NSF)
Institute
Division of Computer and Network Systems (CNS)
Type
Standard Grant (Standard)
Application #
1218181
Program Officer
John Brassil
Project Start
Project End
Budget Start
2012-08-01
Budget End
2016-07-31
Support Year
Fiscal Year
2012
Total Cost
$365,987
Indirect Cost
Name
Rutgers University
Department
Type
DUNS #
City
Newark
State
NJ
Country
United States
Zip Code
07102