via distributed cooperative control application to power management in computing clusters Authors Mianyu Wang Nagarajan Kandasamy Allon Guez and Moshe Kam Proceedings of the 3 ID: 674357
Download Presentation The PPT/PDF document "Adaptive performance control of computin..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Adaptive performance control of computing systems via distributed cooperative control: application to power management in computing clusters
Authors:
Mianyu
Wang,
Nagarajan
Kandasamy
,
Allon
Guez
, and Moshe
Kam
Proceedings of the 3
rd
International Conference on Autonomic Computing, ICAC 2006, Dublin, Ireland
Presenter:
Ramya
Pradhan
,
Fall 2012, UCF.Slide2
OutlineResearch problemProposed solutionEvaluation of proposed solutionStrengths
Limitations
Proposed extensionsSlide3
Research Problem
Server cluster
Clients
Power
Consumption
How to balance power consumption with
time-varying workload and
QoS
?Slide4
Proposed solutionFully decentralized and
cooperative control framework
using optimal control theory
balance
cluster operating frequency
and
average response time
scalable due to problem decompositionfault-tolerant due to cooperative controlno intra-cluster communicationSlide5
Proposed solution using optimal controlOptimal control
“uses
predictive approach
that generates
sequence of control inputs
over a specified
lookahead horizon
while estimating changes in operating conditions.”System ModelSystem state: queue sizeConstrained control input: operating frequency
Output: average response timeSlide6
Distributed control framework
Server cluster
Global request buffer
Clients
Dynamic
ControllersSlide7
EvaluationSystem settingse-commerceVirtual store consisting of 10000 objects
response time uniformly chosen between (4,11)
ms
request distribution
popularity
temporal locality
cluster of four serversSlide8
Evaluation
Adaptive power consumptionSlide9
Evaluation
Adaptive power consumption during processors’ failure Slide10
StrengthsDevelopment of a communication-less framework for distributed optimization
Implementation of the framework
of
power consumption
and
guarantee
QoS
Usage of distributed frameworkautonomous controllersno single point of failurecapable of self-* propertiesSlide11
LimitationsMain concept: decomposing power management into optimal control problems for each server, based on the assumption
that resource provisioning and allocation can also be decomposed into such problems; this may not always be possible.
Adding new servers adds to the overhead in predicting its behavior by all other servers. Results for adding servers is
not
presented.Slide12
Possible extensionsStudy the system under dynamic
adding and removing of servers
Experiment with perturbations when servers are
optimally performing
remove servers that
almost always
guanrantee
QoS and see how other servers respondadd more servers to observe how estimating
the other servers’ behavior affects guarantee of QoSSlide13
Thank You!