Cutting the electric bill for internet-scale systems

By Asfandyar Qureshi, Rick Weber, Hari Balakrishnan, John V. Guttag, Bruce M. Maggs. Conference on Applications, Technologies, Architectures, and Protocols for Computer Communication, 2009

303 citations recorded by the source corpus.

Energy expenses are becoming an increasingly important fraction of data center operating costs. At the same time, the energy expense per unit of computation can vary significantly between two different locations. In this paper, we characterize the variation due to fluctuating electricity prices and argue that existing distributed systems should be able to exploit this variation for significant economic gains. Electricity prices exhibit both temporal and geographic variation, due to regional demand differences, transmission inefficiencies, and generation diversity. Starting with historical electricity prices, for twenty nine locations in the US, and network traffic data collected on Akamai's CDN, we use simulation to quantify the possible economic gains for a realistic workload. Our results imply that existing systems may be able to save millions of dollars a year in electricity costs, by being cognizant of locational computation cost differences.

Subfields: Computer Networks and Communications, Information Systems

View Paper: https://www.semanticscholar.org/paper/95c20aa9ac4cd28bebcab48d0d385631a12d5377

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Information Systems

Citations in the corpus

Predicting the Performance of Virtual Machine Migration (2010, unknown): “and distributed generation from renewable energy sources in addition to exploiting geographic and temporal variation in electricity prices [25].”
The Akamai network: a platform for high-performance internet applications (2010, background): “This also has an environmental cost when underutilized infrastructure consumes significant amounts of power [33].”
It's not easy being green (2012, background): “These inherent complexities have not been addressed in recent work, which focus essentially only on either the electricity cost [30] [35] [37] or the carbon footprint of datacenters [9] [26] [29].”
Greening geographical load balancing (2011, background): “Recently, geographical load balancing has been suggested to reduce energy cost by exploiting the electricity price di.erences across regions.”
Optimal power cost management using stored energy in data centers (2011, background): “Another line of work has looked at cost reduction opportunities offered by geographical variations within utility prices for data centers where portions of workloads could be serviced from one of several locations [11,18].”
Inter-datacenter bulk transfers with netstitcher (2011, background): “For example, in [41] it has been shown that by taking advantage of spatial and temporal variation of energy prices, a datacenter can save millions in its energy bills.”
A Survey of Green Networking Research (2010, unknown): “Companies like Amazon explore such geographical delocalization of services [11] in order to reduce the operational expenditures related to energy supply.”
GreenHadoop: leveraging green energy in data-processing frameworks (2012, background): “Most works that have considered variable energy prices have targeted request distribution across multiple datacenters in interactive Internet services [17, 19, 22, 24].”
A Survey of Large Scale Data Management Approaches in Cloud Environments (2011, background): “hosts? As pointed in [44], electricity costs can vary each hour, motivating the implementation of an algorithm to decide how, in a certain time of the day, the load of the cloud should be geographically distributed.”
Virtual machine power metering and provisioning (2010, background): “Also, electricity price varies every few minutes on the whole sale market and can be exploited by large data centers to reduce operating costs [21].”
Starfish: A Self-tuning System for Big Data Analytics (2011, unknown): “Efficient scheduling of a Hadoop workflow is further complicated by concerns like (a) avoiding cascading reexecution u nder node failure or data corruption [11], (b) ensuring power pro portional computing, and (c) adapting to imbalance in load or co st of energy across geographic regions and time at”
Capping the brown energy consumption of Internet services at low cost (2010, background): “[29] proposed a greedy heuristic to distribute requests across data centers based on electricity prices that vary hourly.”