• Publications
  • Influence
ADAPTLOAD: effective balancing in clustered web servers under transient load conditions
Detailed performance comparisons via simulation using traces from the 1998 World Cup show that ADAPTLOAD is robust as it consistently outperforms traditional load balancing policies, especially under conditions of transient overload.
Workload-aware load balancing for clustered Web servers
It is shown that the proposed load, balancing policy based on statistical information derived from recent workload history provides similar performance benefits as locality-aware allocation schemes, without requiring locality data.
CPU Load Predictions on the Computational Grid *
This paper presents and evaluates a new and innovative method to predict the one-stepahead CPU load in a grid that forecasts the future CPU load based on the tendency in several past steps and in previous similar patterns, and uses a polynomial fitting method.
End-to-end delay bounds for traffic aggregates under guaranteed-rate scheduling algorithms
  • Wei Sun, K. Shin
  • Computer Science
    IEEE/ACM Transactions on Networking
  • 1 October 2005
Overall, aggregate scheduling is shown theoretically to provide bounded e2e delays and practically to provide excellent e2E delay performance and it incurs lower scheduling and state-maintenance overheads at routers than per-flow scheduling.
Differentiated BGP Update Processing for Improved Routing Convergence
Two differentiated update processing (DUP) algorithms are proposed and shown to be very effective for large networks, yielding 30% fewer updates and reducing convergence time by 80%.
Dynamic Task Flow Scheduling for Heterogeneous Distributed Computing: Algorithm and Strategy
This paper proposes a task scheduling method including a scheduling strategy which adapts to the dynamic task flow and a genetic algorithm which can achieve the short completion time of a batch of tasks and shows that this scheduler achieves much better scheduling results than the others.
Predicting Running Time of Grid Tasks based on CPU Load Predictions
The experimental results demonstrate that both the CPU load prediction method and task running time prediction strategy outperform significantly the widely used AR(16) load prediction model and the task running-time prediction method based on this model.
Heuristics and Evaluations of Energy-Aware Task Mapping on Heterogeneous Multiprocessors
  • Wei Sun, T. Sugawara
  • Computer Science
    IEEE International Symposium on Parallel and…
  • 16 May 2011
This paper formulate and study the optimization problem of reducing overall completion time and the total energy consumption, and then some heuristics, which are experimentally evaluated and compared.
Practical task flow scheduling for high throughput computational grid
A scheduling strategy which adapts to the dynamic task flow and a genetic algorithm which balances the loads of the nodes furthest are developed and simulation results show that the scheduler can adapt to the change of arrival rates better than other schedulers.