FLEET: Fine-grained, Lightweight Energy Estimate Tracing for Microservice Architectures

Modern computing infrastructures account for a significant portion of global energy consumption. Despite recent developments in energy efficient hardware and increased use of renewable energy in computing, current predictions indicate that data center energy usage is growing faster than our hardware-centric efficiency solutions can answer. Microservice architectures (MSAs) decompose applications into loosely-coupled, distributed components that interact to provide an overall service. MSAs are widely used in cloud data centers to achieve business objectives. We believe there is a compelling need to develop real-time, request level energy awareness in MSAs to decrease data center energy costs. In this work, we propose FLEET-a lightweight, sidecar-based deployment for real-time, distributed request level energy tracing. We provide experimental results from a prototype deployment of our system with promising energy estimates at the container and request level.

Author

• Meadows, Catherine

Language

• en

Keyword

• Machine learning
• Energy monitoring
• Energy-aware computing
• Energy modeling
• Microservices

Date created

• 2023

Type of Work

• Master's Thesis

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Computer Science

Degree

• M.S.

Advisor

• Wood, Timothy
• Parmer, Gabriel

Committee Member(s)

• Narahari, Bhagirath
• Parmer, Gabriel
• Wood, Timothy

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/5t34sk50c