GridSim: a toolkit for the modeling and simulation of distributed resource management and scheduling for Grid computing
Authors
Discussion
Loading...
Member takes
Implication
Simulation study demonstrates repeatable evaluation of heterogeneous grid scheduling algorithms, highlighting a flexible framework for distributed resource management.
Key Points
- To develop a discrete-event simulation toolkit that enables controllable and repeatable evaluation of distributed resource management and scheduling algorithms across heterogeneous Grid computing environments.
- Developed GridSim, a Java-based discrete-event toolkit supporting modeling of time- and space-shared heterogeneous Grid resources, users, and tasks.
- Provided software primitives for creating tasks, mapping workloads to distributed resources, and managing varying policy and resource conditions.
- Simulated a Nimrod-G-like resource broker to evaluate the performance of deadline- and budget-constrained scheduling algorithms designed for cost- and time-minimization.
- Established a controllable and repeatable simulation platform capable of modeling multi-domain Grid environments with diverse administrative policies.
- Successfully demonstrated performance benchmarking of time- and cost-minimization scheduling algorithms under strict deadline and budget constraints.
Cite This Study
Buyya et al. (2002) studied this question.