edge-computing-testbed is a testbed for task offloading in edge computing. It uses containers to emulate the edge node. It supports schedulers based on deep reinforcement learning.
The testbed needs to install Java 17+.
We recommend installing the K8S via the KubeKey.
https://www.kubesphere.io/docs/v3.3/quick-start/all-in-one-on-linux/
It also will install docker.
build the docker image based on docker/build-image.md for components in the testbed. The components include edge-node, edge-controller and edge-experiment.
Execute kubectl apply -f k8s/mysql.yaml to install MySQL.
Execute sql/create_table.sql to create database and tables in MySQL.
Execute kubectl apply -f k8s/nacos.yaml to install Nacos.
Add all the configurations in nacos/.
Install Hadoop 3.3.4 using Pseudo-Distributed Installation via the official document.
we also provide the installation document in hadoop/hadoop-installation.md.
The testbed consists of three layers and middleware. These three layers include the DRL layer, control layer, and execution layer. The middleware includes Service Registry, Services Configuration, Database, and Distributed File System(DFS).
Start experiment:
start-experiment.shStop experiment:
stop-experiment.sh