Serverless efficient resource allocation method and system based on reinforcement learning

• Main Authors: CHENG WEN, CHEN GUANG, ZENG LINGFANG, LI YONG, ZHANG HUANYU, ZHAO LAIPING
• Format: Patent
• Language: Chinese; English
• Published: 05.05.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a Serverless efficient resource allocation method and system based on reinforcement learning, and the method guarantees a set performance delay target while minimizing the resource configuration consumption of a server-free system through the observation of a tail delay, a decision frequency and a resource efficiency relation. According to the method, the advantage of efficient resource management brought by high-frequency management is fully utilized, the state of each request is observed, and a decision is made for instance resource configuration for processing the request by utilizing a reinforcement learning model. And aiming at the characteristic of multi-stage operation of the function workflow and the lightweight design of the decision model, the high-frequency control layer hides the time overhead and reduces the resource overhead. Compared with the latest workflow task scheduling system, the method has the advantages that the CPU utilization rate is increased, 99% of request d