Security-constrained Workflow Scheduling In Cloud Computing Environments

Cloud computing as an emerging technology, is developing at an explosive way, and down to every corner of our Internet life. Cloud computing with its virtualization technology converting a large amount of physical computing resources into a virtual resource pool, provide on-demand, rapid elasticity and measured services for lots of users. As a cloud platform manages a large amount of computing resources and provides them to a lot of users, the resource scheduling problems in cloud environment has become a hot topic in this area.Recently, most researches on the resource scheduling problems in cloud computing environment, focus on satisfying the customer’s needs of the task’s makespan and cost, and rarely consider the user’s security requirements in the resource scheduling process. To solve this problem, we analyze the security issues in the cloud computing environment and propose a cloud computing security-constrained resource scheduling model. Since safe scheduling policy is overly conservative and the risky scheduling policy is too aggressive, we use a resilient mechanism in this model to meet the user’s security needs. To some extent, it can also reduce the task’s makespan and scheduling costs.Based on this model, we formally defined the security-constrained workflow scheduling problem in Cloud Computing Environments, and use particle swarm optimization(PSO), ant colony algorithm(ACO) and genetic algorithm(GA) to solve it, which are the most frequently used intelligent optimization algorithms in the task scheduling problems.In order to avoid trapping into local best, and improve the solution accuracy, the variable neighborhood search heuristics is introduced into the PSO. We give the schedule strategy of the Variable Neighborhood Particle Swarm Optimization(VNPSO) in detailed steps. Two representative meta-heuristic based scheduling algorithms including Max-Min Ant Colony Optimization(MMACO) and Adaptive Genetic algorithm(AGA) are also implemented as the candidate scheduling strategy.Finally, CloudSim is set up as the simulation platform. Comparing these three scheduling strategies, the simulation experiments show that the variable neighborhood search heuristic provides a good balance between global exploration and local exploitation and makes VNPSO feasible and effective for scheduling workflow tasks.