Multi‐DAGs Scheduling Integrating with Security and Availability in Cloud Environment

Abstract

In view of the issue concerns multiple Di- rected acyclic graphs (DAGs) scheduling in multi-tenant cloud computing environment, a scheduling strategy that integrate security and availability is proposed to satisfy the tenants' requirements for resource security and avail- ability, as thus it can not only protect the users' privacy and data security but also advance the success rate. The proposal assesses resource reputation to ensure jobs can be scheduled onto relatively security nodes; during task scheduling, it classifies the DAGs to achieve fairness; in the process of resources allocation, the objective function would maximize the user's security satisfaction and mini- mize the deviation of availability; meanwhile, it takes ad- vantage of time chips flexibly to promote resource uti- lization rate; afterwards, we present a Greedy algorithm integrating with security and availability (GISA) to im- plement the strategy. The experimental results show the correctness and superior of the novel strategy.