lorenzofelletti/EdgeCloudApproximate

EdgeApproximate

Environment Setup

• Install Docker Desktop
• Run docker-compose up -d in the root directory of the project
• Enjoy!

Containers Configuration

• There are 5 containers in total
  • One Spark Master
  • One Spark Worker
  • One Zookeeper node
  • Two Kafka nodes
    • Kafka (Main Kafka broker)
    • Edge1 (Edge node).

On the Spark master node, runs a pyspark script that continuously reads from the DATAOUT topic(s) of the Kafka broker and writes the processed data to a local directory. Also on the Spark master node, runs a Flask server that serves the processed data to the frontend. The Flask server is accessible at localhost:5000.

On the Zookeeper node, runs a Zookeeper server (required by Kafka).

On the Kafka node, two programs runs:

• kafka-producer-rs -- A Rust program that reads from a local csv the data to distribute to the Kafka nodes via a DATAIN topic; simulating the data arriving to the edge nodes.
• kafka-edge-rs -- A Rust program that reads from the DATAIN topic and writes to one or more DATAOUT topics (depending on the configuration); simulating being an edge node.

On the Edge1 node, runs the same kafka-edge-rs program as the Kafka node, but withouth the kafka-producer-rs program (required only on one node).