EDC Connector

Built with

One of the guiding principles in developing the connector is simplicity and keeping the core small and efficient with as little external dependencies as possible to avoid version conflicts. We do not want to force any third-party dependencies onto our users, so we aim to avoid any of the big frameworks. Of course, if you want to use them, you still can add them to your extensions (see: [TBW]). The connector is a plain Java application built with Gradle, but it can be embedded into any form of application deployment.

Documentation

Developer documentation can be found under docs/developer, where the main concepts and decisions are captured as decision records.

Some more documentation can be found at extensions, launchers and the samples repository.

For detailed information about the whole project, please take a look at our GitHub pages.

Getting Started

Onboarding and first steps

If you are not yet familiar with the EDC nomenclature, we strongly advise reading the documentation and/or watching the introductory videos.

Samples

The Samples repository provides a set of examples for getting familiar with the EDC framework and learning how to set up and use a connector. The samples begin with the very basics (e.g. learning how to write an extension or use a configuration file) and then move on to more complex scenarios (e.g. performing different data transfers). More samples will be added in the future, so be sure to check back regularly.

Add Maven dependencies

Official versions are available through MavenCentral . Please add the following instructions in your build.gradle[.kts] file (if not already present):

repositories {
    mavenCentral()
    // ... other maven repos
}

We strongly recommend to use official versions and only switch to snapshots if there is a clear need to do so, or you've been instructed to do so, e.g. to verify a bugfix.

All artifacts are under the org.eclipse.edc group id, for example:

dependencies {
    implementation("org.eclipse.edc:spi:core-spi:<<version>>")
    // any other dependencies
}

Using SNAPSHOT versions

In addition, EDC regularly publishes snapshot versions, which are available at Sonatype's snapshot repository. In order to add them to your build configuration, simply add this:

repositories {
    mavenCentral()
    maven {
        url = uri("https://oss.sonatype.org/content/repositories/snapshots/")
    }
    // any other repos
}

Then you can add snapshot dependencies by simply using the -SNAPSHOT version suffix:

dependencies {
    implementation("org.eclipse.edc:spi:core-spi:0.4.2-SNAPSHOT")
    // any other dependencies
}

You may check MavenCentral for a comprehensive list of all official versions.

Please be aware of the following pitfalls:

• snapshots are by definition unstable - every new snapshot replaces an old one
• this may cause unrepeatable builds
• snapshots are created irregularly, we do not have any fixed publish schedule

Using release versions

We plan to have actual release versions starting some time mid 2022. Please check back soon.

For more information about versioning please refer to the release documentation

Checkout and build from source

The project requires JDK 11+. To get started:

git clone [email protected]:eclipse-edc/Connector.git

cd Connector

./gradlew clean build

That will build the connector and run unit tests.

[Optional] Setup your IDE

If you wish to configure your IDE/editor to automatically apply the EDC code style, please follow this guide.

Note: the style guide will be checked/enforced in GitHub Actions.

Generate the OpenApi specification

Please refer to this document.

Directory structure

spi

This is the primary extension point for the connector. It contains all necessary interfaces that need to be implemented as well as essential model classes and enums. Basically, the spi modules defines the extent to what users can customize and extend the code.

core

Contains all absolutely essential building that is necessary to run a connector such as TransferProcessManager, ProvisionManager, DataFlowManager, various model classes, the protocol engine and the policy piece. While it is possible to build a connector with just the code from the core module, it will have very limited capabilities to communicate and to interact with a data space.

extensions

This contains code that extends the connector's core functionality with technology- or cloud-provider-specific code. For example a transfer process store based on Azure CosmosDB, a secure vault based on Azure KeyVault, etc. This is where technology- and cloud-specific implementations should go.

If someone were to create a configuration service based on Postgres, then the implementation should go into the extensions/database/configuration-postgres module.

launchers

Launchers are essentially connector packages that are runnable. What modules get included in the build (and thus: what capabilities a connector has) is defined by the build.gradle.kts file inside the launcher subdirectory. That's also where a Java class containing a main method should go. We will call that class a "runtime" and in order for the connector to become operational the runtime needs to perform several important tasks (="bootstrapping"). For an example take a look at this runtime

data-protocols

Contains implementations for communication protocols a connector might use, such as DSP.

Releases

GitHub releases are listed here. Please find more information about releases in our release approach.

Roadmap

See here for more information about project and milestone planning. Scheduled and ongoing milestones are listed here.

Tags

Available tags can be found here.

Contributing

See how to contribute.