Privacy and security-aware workflow scheduling in a hybrid cloud

Abstract

Increasing numbers of workflow applications are being deployed on cloud platforms. When deploying workflows in a hybrid cloud environment, additional concerns emerge: 1. privacy-sensitive data and tasks cannot be exposed to a public cloud platform; 2. security and bandwidth of data transmission across cloud platforms should be guaranteed as it goes through the ‘unpredictable’ Internet. This paper investigates how to effectively reduce monetary cost when deploying a workflow in a hybrid cloud under the deadline and privacy constraints. A three-level data privacy and security model integrating hybrid encryption is devised to guarantee privacy and security of workflow applications. Two scheduling heuristics are then proposed to tackle the considered constrained optimization problem: privacy and security-aware list scheduling (PSLS) and simulated annealing (PSSA). PSLS distributes the user-defined deadline to each task and allocates each task to a resource in the hybrid cloud that meets constraints and incurs a minimum cost. PSSA permutes task lists for iterative improvement based on simulated annealing and PSLS. Simulation experiments are conducted and experimental results demonstrate that PSLS generally performs better than four existing algorithms to optimize monetary costs and PSSA achieves an even higher performance at the expense of runtime.