Open Package Deploy

OPD (Open Package Deploy) allows you to deploy almost everything, everywhere, easily.

The code for the backend (server and agent) and the frontend is available on GitHub.

You can find some screenshots of the UI in the User Guide or this image folder.

• Open Package Deploy
• Features
• What can be deployed
• Installation
• User Guide
• Agent Functionality
  • When will the agent contact the Server
  • Update Request Content
• Security
  • Agent-Server Communication
  • Packages
  • Agent Registration
• Privacy

Features

• Simple
• Fast
• Cross-Platform Agent (macOS, Linux, Windows)
• Secure (agent - server communication is fully encrypted and signed)
• Multi-user capable
• Web app is mobile-friendly
• Agent Auto-Update

What can be deployed

OPD was designed to be usable for many different use cases.

The agent will execute a script (start.sh for Linux and macOS and start.bat for Windows) that you provide inside a package. From there, you can do almost anything.

You can get more info in the Guide: Create a Package

Examples:

• Install software through a package manager or an included installer file
• Copy included files to a destination
• Change system configurations

Example packages will be provided in this GitHub repository.

Installation

To get OPD up and running, follow the Quickstart Guide. There, you will set up OPD, Keycloak and the required databases using Docker and Docker Compose.

Don't forget to check out the Recommended Keycloak Settings

User Guide

If you want to learn more about how to use OPD, head over to the User Guide.

Agent Functionality

This is a short description of how and when the agent will contact the server and what information is transmitted.

When will the agent contact the Server

The agent will contact and ask the server for updates in three scenarios:

• Agent Startup
• Update Interval timer is expired
• After successfully processing a Package

Update Request Content

When the agent asks the server for updates, it will get and process the following data:

• Agent Checksum (for Auto Update): Start the auto-update if the checksum is different
• Update Interval: Update the value in the local config file and restart the Agent
• Is a deployment available: Start the package download and installation process

Security

Agent-Server Communication

The communication between the agent (endpoint/client) and the server is fully encrypted and signed.

Every request is first signed and then encrypted using asymmetric cryptography. The signature uses the SHA512withECDSA-algorithm, while the encryption uses ECIES/None/NoPadding.

If a request is received, the content is first decrypted and then verified with the signature. If an error occurs, the request gets dropped.

Every agent will generate its own public- and private-key pair. During the registration process, the agent will send its public key to the server and get the servers public key.

Packages

All uploaded packages will be encrypted before they can be delivered to agents. A package will never be sent in plaintext to an agent.

Every package is encrypted using a unique AES-key and IV for the AES/GCM/NoPadding-Algorithm.

After the package is uploaded, checksums will be calculated, and the package will be encrypted. The plaintext package will be deleted afterward.

Agent Registration

To initiate the registration process, a Registration Token needs to be sent. This token is auto-generated on first start and can be seen and updated/regenerated in the Settings.

Privacy

OPD will only communicate with its own services (in the docker-compose.yml).

OPD has no tracking or analytics features. All data is stored on your system, and nothing is shared.