Cloud data centers are characterized by both high energy consumption and low resource utilization. To meet the peak load, cloud data centers generally accept over-provisioning of computing resources and only 10-50 percent of their full capacity is utilized. Furthermore, even though the computing nodes are underutilized, they consume a significant amount of power. For example, a server that utilizes 10 percent of its computing capacity can consume over 50 percent of its peak power. Therefore, consolidation, which collects the workload into a reduced number of computing nodes and make the idle nodes sleep, is one of the most important techniques for cloud data centers to improve their resource utilization and energy efficiency. Since our proposed technique aims to feasibly run DAG-based real-time tasks on the minimum number of processors, it can be used as a basic building block for designing efficient consolidation techniques for cloud data centers.

[Related Publications]

Hyeonjoong Cho, Chulgoo Kim, Joohyung Sun, Arvind Easwaran ,Juderk Park, Byeongcheol Choi, "Scheduling Parallel Real-time Tasks on the Minimum Number of Processors", IEEE Transactions on Parallel and Distributed Systems, 2019.

Joohyung Sun, Hyeonjoong Cho, A. Easwaran, J. Park, B. Choi, "Flow Network-Based Real-Time Scheduling for Reducing Static Energy Consumption on Multiprocessors", IEEE Access, 7, 1330-1344.

Joohyung Sun, Hyeonjoong Cho, "Power-efficient real-time scheduling based on multi-granularity resource reservation for multimedia services", IET Software, Volume 11, Issue 4, August 2017, p. 171 – 180