Context for scalable application

A distributed way to share variables in a scalable environment

Description

This context server is an RPC server giving access to a shared dictionary that is meant to hold variables and offer strategies in a scalable way.

Installing

Native installation

Starting by a preconfigured rabbitMQ server:

# Giving your rabbit_mq variables
export RABBIT_MQ_HOST="localhost"
export RABBIT_MQ_USER=""
export RABBIT_MQ_PASSWORD=""

Configure your application's context rpc server:

export CONTEXT_RPC_PORT="localhost"
export CONTEXT_RPC_HOST="6060"

Now, this variable is meant to be handled by another system, use the value 1 as a start:

# This one is a specific ID given by another subsystem
export ZOS_CONTEXT_ID="1"

Docker

Use this command to build and deploy the containers:

sudo docker-compose up -d

Hypothesis

We suppose that the beholding nodes of this shared context have a way of knowing each other. The subsystem responsible for making knowledge between the nodes is meant keep groups of nodes knowing eachother, another feature that I plan to add is the election process using the sidecar pattern to elect one coordinator from within each group in order to communicate with other groups. But for now, I just suppose that these subsystems are already made.

Progress

• Current code consistency.
• Consensus handling.
  • In between node communication.
  • Stateless strategy instances.
    • Key value store. (REDIS)
• Accessibility.
  • Context as an RPC service.
  • Strategies:
    • Scheduling:
      • Round robin
  • Text storage.
• Containerisation. (Dockerfile)
  • Automation of deployment.
  • Smaller footprint.