PoMic: Dynamic Power Management of VM-Microservices in Overcommitted Cloud

Changes in user requests and data processing volume induce changes in the processing pattern and demand for computing resources. One of the new models used in clouds is a microservices architecture. In the microservice model, each application consists of loosely coupled services. Global large-scale data centers are expanding due to the introduction of new technologies such as microservice in clouds. Electricity consumption is a crucial issue in data centers. However, electricity sources emit a significant amount of carbon dioxide into the environment. This paper proposes a novel method for managing the dynamic power consumption of microservices in cloud data centers. This approach assumes microservices located on virtual machines and follows a decision process to consolidate VM-microservices based on migration or resized virtual machines. The method aims to increase the productivity of computing resources and satisfy the Service Level Agreement (SLA) for the respective services. Moreover, the approach was evaluated on the PlanetLab dataset on the CloudSimPlus platform. The results showed that using the decision process reduced energy consumption by at least 10%.

Automated summary

Changes in user requests and data processing volume lead to changes in processing pattern and demand for computing resources. The introduction of microservices architecture in clouds has resulted in the expansion of global large-scale data centers. However, the electricity consumption of these data centers and its environmental impact is a crucial issue. This paper proposes a novel method for managing the dynamic power consumption of microservices in cloud data centers, which aims to increase computing resource productivity and satisfy Service Level Agreements.