Cloud scheduling improves energy use and container efficiency with graph learning
Cloud Workflow Scheduling Based on Graph Attention-Driven Hierarchical Reinforcement Learning
Machine LearningNetworking and Internet Architecture
Summary
Scheduling tasks in cloud computing is tricky because tasks depend on each other and use shared resources. The authors created a new method that uses graph attention networks to better understand task dependencies and deadlines. They combined this with a two-part reinforcement learning system that decides task placement and container allocation. Their approach balances completing tasks on time with using resources efficiently and saving energy. Tests on real cloud data showed their method can keep success rates high while lowering energy use and increasing container usage.
What this means in practice
- •For cloud infrastructure engineers: Enhance task scheduling in cloud platforms to improve resource use and reduce energy consumption while meeting deadlines.
- •For data center operators: Optimize container placement and management to raise efficiency and lower operational costs through smarter scheduling.
Authors
Zongjin Li, Shaohan Feng, Chunxi Yang, Wenbo Wang
Abstract
Dynamic cloud workflow scheduling must balance deadline satisfaction, container utilization, and energy consumption while dealing with stochastic task-execution speeds, placement-dependent communication, and coupled task and container decisions. Workflows are naturally modeled as directed acyclic graphs (DAGs), but conventional vector- or matrix-based states do not fully capture their dependency topology. To better represent task urgency and structural relationships, we assign predicted sub-deadlines to tasks and use a multi-head graph attention network (GAT) to extract dependency information from the evolving DAGs. Based on these representations, we develop a Graph Attention-Driven Hierarchical Reinforcement Learning (GA-HRL) framework and model the scheduling process as an event-driven hierarchical semi-Markov decision process (SMDP). Workflow arrivals and task completions trigger scheduling events. At each scheduling event, the Task Scheduling (TS) agent first processes the currently ready tasks by assigning them to admissible existing containers or requesting new ones. The requested containers are then processed by the Container Scheduling (CS) agent for host placement before the environment advances. The two agents are trained alternately using separate Proximal Policy Optimization (PPO). Experiments on the 2018 Alibaba cluster trace show that GA-HRL maintains competitive workflow success rate and, in settings where success is comparable, generally achieves higher container utilization and lower energy consumption. Under the largest speed variation, it trades a small success-rate margin for substantially lower energy. Simulation code is available at: https://github.com/zongjin130/GA-HRL.