Task Assignment in a Distributed System: Improving Performance by Unbalancing Load
Task Assignment in a Distributed System: Improving Performance by Unbalancing Load
Optimal stochastic scheduling in multiclass parallel queues
SIGMETRICS '99 Proceedings of the 1999 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Hierarchical resource management for Web server clusters with dynamic content
SIGMETRICS '99 Proceedings of the 1999 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Scheduling optimization for resource-intensive Web requests on server clusters
Proceedings of the eleventh annual ACM symposium on Parallel algorithms and architectures
Interpreting Stale Load Information
IEEE Transactions on Parallel and Distributed Systems
On the origin of power laws in Internet topologies
ACM SIGCOMM Computer Communication Review
Impact of Workload and System Parameters on Next Generation Cluster Scheduling Mechanisms
IEEE Transactions on Parallel and Distributed Systems
Task assignment with unknown duration
Journal of the ACM (JACM)
Content management on server farm with layer-7 routing
Proceedings of the 2002 ACM symposium on Applied computing
Stability of load balancing algorithms in dynamic adversarial systems
STOC '02 Proceedings of the thiry-fourth annual ACM symposium on Theory of computing
Performance Evaluation with Heavy Tailed Distributions
JSSPP '01 Revised Papers from the 7th International Workshop on Job Scheduling Strategies for Parallel Processing
Performance Evaluation with Heavy Tailed Distributions
TOOLS '00 Proceedings of the 11th International Conference on Computer Performance Evaluation: Modelling Techniques and Tools
The effectiveness of request redirection on CDN robustness
ACM SIGOPS Operating Systems Review - OSDI '02: Proceedings of the 5th symposium on Operating systems design and implementation
Optimizing Static Job Scheduling in a Network of Heterogeneous Computers
ICPP '00 Proceedings of the Proceedings of the 2000 International Conference on Parallel Processing
A Single System Image Server Cluster using Duplicated MAC and IP Addresses
LCN '01 Proceedings of the 26th Annual IEEE Conference on Local Computer Networks
A dynamic load distribution strategy for systems under high task variation and heavy traffic
Proceedings of the 2003 ACM symposium on Applied computing
The effectiveness of request redirection on CDN robustness
OSDI '02 Proceedings of the 5th symposium on Operating systems design and implementationCopyright restrictions prevent ACM from being able to make the PDFs for this conference available for downloading
Robust Processing Rate Allocation for Proportional Slowdown Differentiation on Internet Servers
IEEE Transactions on Computers
A least flow-time first load sharing approach for distributed server farm
Journal of Parallel and Distributed Computing
Optimal state-free, size-aware dispatching for heterogeneous M/G/-type systems
Performance Evaluation - Performance 2005
Task assignment with work-conserving migration
Parallel Computing
System support for web hosting services on server clusters
Computers and Electrical Engineering
Comparing job allocation schemes where service demand is unknown
Journal of Computer and System Sciences
Analysis of size interval task assignment policies
ACM SIGMETRICS Performance Evaluation Review
A hybrid load balancing strategy of sequential tasks for grid computing environments
Future Generation Computer Systems
Performance Evaluation
Queueing Systems: Theory and Applications
WEA'03 Proceedings of the 2nd international conference on Experimental and efficient algorithms
Microprocessors & Microsystems
To balance or unbalance load in size-interval task allocation
Probability in the Engineering and Informational Sciences
Improving MapReduce performance in heterogeneous environments
OSDI'08 Proceedings of the 8th USENIX conference on Operating systems design and implementation
ACSC '09 Proceedings of the Thirty-Second Australasian Conference on Computer Science - Volume 91
Modelling job allocation where service duration is unknown
IPDPS'06 Proceedings of the 20th international conference on Parallel and distributed processing
Review: Task assignment policies in distributed server systems: A survey
Journal of Network and Computer Applications
M/M/1-PS queue and size-aware task assignment
Performance Evaluation
Dispatching problem with fixed size jobs and processor sharing discipline
Proceedings of the 23rd International Teletraffic Congress
Task assignment based on prioritising traffic flows
OPODIS'04 Proceedings of the 8th international conference on Principles of Distributed Systems
On fairness, optimal download performance and proportional replication in peer-to-peer networks
NETWORKING'05 Proceedings of the 4th IFIP-TC6 international conference on Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; Mobile and Wireless Communication Systems
Minimizing slowdown in heterogeneous size-aware dispatching systems
Proceedings of the 12th ACM SIGMETRICS/PERFORMANCE joint international conference on Measurement and Modeling of Computer Systems
Lookahead actions in dispatching to parallel queues
Performance Evaluation
Hi-index | 0.00 |
We consider the problem of task assignment in a distributed system (such as a distributed Web server) in which task sizes are drawn from a heavy-tailed distribution. Many task assignment algorithms are based on the heuristic that balancing the load at the server hosts will result in optimal performance. We show this conventional wisdom is less true when the task size distribution is heavy-tailed (as is the case for Web file sizes). We introduce a new task assignment policy, called Size Interval Task Assignment with Variable Load (SITA-V). SITA-V purposely operates the server hosts at different loads, and directs smaller tasks to the lighter-loaded hosts. The result is that SITA-V provably decreases the mean task slowdown by significant factors (up to 1000 or more) where the more heavy-tailed the workload, the greater the improvement factor. We evaluate the tradeoff between improvement in slowdown and increase in waiting time in a system using SITA-V, and show conditions under which SITA-V represents a particularly appealing policy.