How eclipse-edc/minimumviabledataspace works
Main system components
• Connector Runtimes:
– Consumer Connector (controlplane and associated dataplane)
– Two Provider Connectors (provider-qna and provider-manufacturing; each has its controlplane and dataplane)
• Catalog Server: A specialized runtime that aggregates catalog pointers from provider connectors.
• IdentityHub Services: One for the consumer and one for the provider; these are responsible for managing DID documents, verifiable credentials, and participant identity data.
• Issuer Service / NGINX: A separate issuer entity that hosts a static DID document (served via an NGINX container) and supplies public key information for signing credentials.
• Data Persistence and Vault Components (in Kubernetes deployments):
– PostgreSQL databases
– HashiCorp Vault instances (for secret management)
• Infrastructure & Deployment:
– Terraform scripts and Kubernetes manifests (e.g. Kind configs, ingress, modules for connectors, catalog server, identityhub, issuer)
– Seed scripts that bootstrap credentials and participant context
Out the relationships and interactions
• Interaction between the Consumer Connector and Provider Connectors via data exchange protocols (noting that they communicate using dataspace protocol HTTP endpoints).
• That the Provider Connectors do not expose their catalogs directly; instead, they register their assets (catalog assets) with the Catalog Server by acting as pointers.
• IdentityHub components: both consumer and provider IdentityHubs are used for obtaining and seeding verifiable credentials (e.g. during contract negotiations, a consumer must present MembershipCredential and a DataProcessorCredential).
• NGINX host for the issuer’s DID document; indicate that its endpoint must be reachable by other components and that the issuer public/private key pair is managed externally using OpenSSL (with re-generation processes shown in the advanced topics).
Architectural patterns and design principles
• The project adopts a modular architecture with microservice-like runtimes (several “launcher” modules) each responsible for a specific concern.
• It follows a service-oriented approach where runtime modules communicate via RESTful APIs (look for HTTP endpoints and Postman collections in the README).
• There is a clear separation between core functionality (credential handling, authentication, contract negotiation), infrastructure concerns (deployment with Terraform, Kubernetes configuration), and custom extension points (see the extensions directory for implementations such as Catalog Node Resolver and custom policy evaluation functions).
Relevant technologies and frameworks
• Gradle is used as the build system (see build.gradle.kts files).
• Java is the primary language (multiple src/main/java directories).
• Docker (for containerizing runtimes and NGINX) and Kubernetes (with Kind and Terraform for deployment) are involved.
• Additional key artifacts include configuration files for Checkstyle, GitHub workflows, and a suite of integration and performance tests.
• Use clear labels for every component (e.g., “Consumer Connector (Controlplane/Dataplane)”, “Provider Connector - Q&A”, “Provider Connector - Manufacturing”, “Catalog Server”, “IdentityHub (Consumer)”, “IdentityHub (Provider)”, “Issuer (NGINX)”, “PostgreSQL”, “Vault”, “Terraform/K8S Infra”)
• Architectural notes, such as the fact that provider connectors publish catalogs indirectly (via the catalog server) and that all DSP requests require additional verifiable credentials.