Power-aware QoS management in web servers

Vivek Sharma, Arun Thomas, Tarek Abdelzaher, Kevin Skadron, Zhijian Lu

Research output: Contribution to conferencePaper

Abstract

Power management in data centers has become an increasingly important concern. Large server installations are designed to handle peak load, which may be significantly larger than in off-peak conditions. The increasing cost of energy consumption and cooling incurred in farms of high-performance web servers make low-power operation during off-peak hours desirable. This paper investigates adaptive algorithms for dynamic voltage scaling in QoS-enabled web servers to minimize energy consumption subject to service delay constraints. We implement these algorithms inside the Linux kernel. The instrumented kernel supports multiple client clases with per-class deadlines. Energy consumption is minimized by using a feedback loop that regulates frequency and voltage levels to keep the synthetic utilization around the aperiodic schedulability bound derived in an earlier publication. Enforcing the bound ensures that deadlines are met. Our evaluation of an Apache server running on the modified Linux kernel shows that non-trivial off-peak energy savings are possible without sacrificing timeliness.

Original languageEnglish (US)
Pages63-73
Number of pages11
StatePublished - Dec 1 2003
Externally publishedYes
Event24th IEEE International Real-Time Systems Symposium RTSS 2003 - Cancun, Mexico
Duration: Dec 3 2003Dec 5 2003

Other

Other24th IEEE International Real-Time Systems Symposium RTSS 2003
CountryMexico
CityCancun
Period12/3/0312/5/03

Fingerprint

Quality of service
Servers
Energy utilization
Adaptive algorithms
Farms
Energy conservation
Cooling
Feedback
Electric potential
Costs
Linux

Keywords

  • Aperiodic task scheduling
  • Dynamic voltage scaling
  • Linux
  • Utilization bounds
  • Web servers

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

Sharma, V., Thomas, A., Abdelzaher, T., Skadron, K., & Lu, Z. (2003). Power-aware QoS management in web servers. 63-73. Paper presented at 24th IEEE International Real-Time Systems Symposium RTSS 2003, Cancun, Mexico.

Power-aware QoS management in web servers. / Sharma, Vivek; Thomas, Arun; Abdelzaher, Tarek; Skadron, Kevin; Lu, Zhijian.

2003. 63-73 Paper presented at 24th IEEE International Real-Time Systems Symposium RTSS 2003, Cancun, Mexico.

Research output: Contribution to conferencePaper

Sharma, V, Thomas, A, Abdelzaher, T, Skadron, K & Lu, Z 2003, 'Power-aware QoS management in web servers' Paper presented at 24th IEEE International Real-Time Systems Symposium RTSS 2003, Cancun, Mexico, 12/3/03 - 12/5/03, pp. 63-73.
Sharma V, Thomas A, Abdelzaher T, Skadron K, Lu Z. Power-aware QoS management in web servers. 2003. Paper presented at 24th IEEE International Real-Time Systems Symposium RTSS 2003, Cancun, Mexico.
Sharma, Vivek ; Thomas, Arun ; Abdelzaher, Tarek ; Skadron, Kevin ; Lu, Zhijian. / Power-aware QoS management in web servers. Paper presented at 24th IEEE International Real-Time Systems Symposium RTSS 2003, Cancun, Mexico.11 p.
@conference{30cebaf952e4454f935edc2f8f79d265,
title = "Power-aware QoS management in web servers",
abstract = "Power management in data centers has become an increasingly important concern. Large server installations are designed to handle peak load, which may be significantly larger than in off-peak conditions. The increasing cost of energy consumption and cooling incurred in farms of high-performance web servers make low-power operation during off-peak hours desirable. This paper investigates adaptive algorithms for dynamic voltage scaling in QoS-enabled web servers to minimize energy consumption subject to service delay constraints. We implement these algorithms inside the Linux kernel. The instrumented kernel supports multiple client clases with per-class deadlines. Energy consumption is minimized by using a feedback loop that regulates frequency and voltage levels to keep the synthetic utilization around the aperiodic schedulability bound derived in an earlier publication. Enforcing the bound ensures that deadlines are met. Our evaluation of an Apache server running on the modified Linux kernel shows that non-trivial off-peak energy savings are possible without sacrificing timeliness.",
keywords = "Aperiodic task scheduling, Dynamic voltage scaling, Linux, Utilization bounds, Web servers",
author = "Vivek Sharma and Arun Thomas and Tarek Abdelzaher and Kevin Skadron and Zhijian Lu",
year = "2003",
month = "12",
day = "1",
language = "English (US)",
pages = "63--73",
note = "24th IEEE International Real-Time Systems Symposium RTSS 2003 ; Conference date: 03-12-2003 Through 05-12-2003",

}

TY - CONF

T1 - Power-aware QoS management in web servers

AU - Sharma, Vivek

AU - Thomas, Arun

AU - Abdelzaher, Tarek

AU - Skadron, Kevin

AU - Lu, Zhijian

PY - 2003/12/1

Y1 - 2003/12/1

N2 - Power management in data centers has become an increasingly important concern. Large server installations are designed to handle peak load, which may be significantly larger than in off-peak conditions. The increasing cost of energy consumption and cooling incurred in farms of high-performance web servers make low-power operation during off-peak hours desirable. This paper investigates adaptive algorithms for dynamic voltage scaling in QoS-enabled web servers to minimize energy consumption subject to service delay constraints. We implement these algorithms inside the Linux kernel. The instrumented kernel supports multiple client clases with per-class deadlines. Energy consumption is minimized by using a feedback loop that regulates frequency and voltage levels to keep the synthetic utilization around the aperiodic schedulability bound derived in an earlier publication. Enforcing the bound ensures that deadlines are met. Our evaluation of an Apache server running on the modified Linux kernel shows that non-trivial off-peak energy savings are possible without sacrificing timeliness.

AB - Power management in data centers has become an increasingly important concern. Large server installations are designed to handle peak load, which may be significantly larger than in off-peak conditions. The increasing cost of energy consumption and cooling incurred in farms of high-performance web servers make low-power operation during off-peak hours desirable. This paper investigates adaptive algorithms for dynamic voltage scaling in QoS-enabled web servers to minimize energy consumption subject to service delay constraints. We implement these algorithms inside the Linux kernel. The instrumented kernel supports multiple client clases with per-class deadlines. Energy consumption is minimized by using a feedback loop that regulates frequency and voltage levels to keep the synthetic utilization around the aperiodic schedulability bound derived in an earlier publication. Enforcing the bound ensures that deadlines are met. Our evaluation of an Apache server running on the modified Linux kernel shows that non-trivial off-peak energy savings are possible without sacrificing timeliness.

KW - Aperiodic task scheduling

KW - Dynamic voltage scaling

KW - Linux

KW - Utilization bounds

KW - Web servers

UR - http://www.scopus.com/inward/record.url?scp=0348195815&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0348195815&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0348195815

SP - 63

EP - 73

ER -