Skip to content

Latest commit

 

History

History
40 lines (31 loc) · 1.43 KB

README.md

File metadata and controls

40 lines (31 loc) · 1.43 KB
Raw

LatenSeer

The repo for SoCC23 Paper: LatenSeer: Causal Modeling of End-to-End Latency Distributions by Harnessing Distributed Tracing.

diagram

What is LatenSeer

LatenSeer is a modeling framework for estimating end-to-end latency distributions in microservice-based web applications.

  • An offline tool.
  • Harnesses distributed tracing.
  • Enables what-if analysis, predicting the potential impacts on end-to-end latency distribution due to various changes in service latencies.

Repo Structure

The repo includes the source code of LatenSeer, scripts for setting up DSB experiments, and a script for collecting DSB Jaeger traces.

How to use LatenSeer

First, you can create a virtual environment and activate it:

python -m venv .venv
source .venv/bin/activate

Then install LatenSeer:

pip install .

Now you are good to do a test run. We provide a simple tutorial about how to use LatenSeer.

Citation

@inproceedings{zhang2023-latenseer,
  title={LatenSeer: Causal Modeling of End-to-End Latency Distributions by Harnessing Distributed Tracing},
  author={Zhang, Yazhuo and Isaacs, Rebecca and Yue, Yao and Yang, Juncheng and Zhang, Lei and Vigfusson, Ymir},
  booktitle={ACM Symposium on Cloud Computing (SoCC’23)},
  year={2023}
}