This repository contains scripts and utilities to automate the deployment of Serverledge.

The repository mostly consists of Ansible playbooks and roles, which can be used to deploy Serverledge over existing hosts. In addition, we also provide Terraform scripts in the terraform directory, which can be used to provision a set of Amazon EC2 instances.

Currently, we only support deployment to Debian-based Linux distributions or to Amazon Linux 2.

Important note: while sensible defaults are used for most the variables, you should check and customize the scripts we provide before using them.

Requirements

Install the following Ansible role, which is used to configure Docker:

ansible-galaxy install geerlingguy.docker

Infrastructure Setup on AWS

We currently provide scripts to reproduce the following configuration:

• AWS-based infrastructure
• 1+ Edge node(s) in a given AWS region
• 1+ Cloud node(s) in a given AWS region
• 1 Load Balancer deployed in one of the Cloud nodes
• 1 Etcd server deployed in one of the Cloud nodes
• (Optional) 1 Client node deployed alongside the Edge node(s)

We assume that the AWS CLI/SDK has already been configured with the required credentials on the machine.

Enter the terraform/ directory and run terraform init.

Edit the file vars.tfvars as needed. Use this file to define how many Edge/Cloud/Client nodes you desire.

Configure the infrastructure:

make tf

When you are done, clean up:

make destroy

Ansible

Inventory

Create an Ansible inventory file (default path is inventory/default.ini). Group the hosts using the following groups:

• sedge_Edge: Edge nodes
• sedge_Cloud: Cloud nodes
• sedge_Client: node(s) acting as client/workload generator

If you provisioned the infrastructure in EC2, you can also use inventory.aws_ec2.yaml that allows Ansible to dynamically retrieve information about our EC2 instances. Update inventory.aws_ec2.yaml (if needed) to set the regions where EC2 instances have been launched.

You can check that everything works by running:

ansible-inventory -i <inventory_file> --list

Deployment

As Serverledge is currently under development, we deploy locally built binaries. To this end, you must provide the local path to Serverledge binaries updating the serverledge_local_bin_dir variable (either modifying the relevant file or providing Ansible "local vars").

Run the playbook:

ansible-playbook [-i <inventory_file>] deploy_serverledge.yml

Run a simple check to verify that the cluster is up:

ansible-playbook [-i <inventory_file>] check.yml

Benchmarking with JMeter

ansible-playbook [-i <inventory_file>] run_experiment.yml