Scheduling Security-Sensitive Task-Chains in Hybrid Cloud Environments

Abstract

The growing adoption of Internet of Things (IoT) devices generates massive data volumes that increasingly strain centralized cloud infrastructures, particularly for latency-sensitive and privacy-aware applications. Hybrid cloud architectures mitigate these limitations by enabling private clouds to dynamically leverage public cloud resources. However, workloads in such environments often include jobs with heterogeneous privacy requirements, making security-aware scheduling essential for maintaining Quality of Service (QoS). This paper investigates scheduling techniques for applications with chained tasks in hybrid cloud systems. Privacy-sensitive jobs (private jobs ) are restricted to private cloud resources, whereas less sensitive jobs (cloud jobs) may execute on either private or public infrastructures. To improve the responsiveness of private jobs, we propose an imprecise-computation–based scheduling approach for cloud jobs and evaluate its effectiveness through simulation under diverse workload conditions.