Technical Deep Dive EDC

📅 June 30, 2025 | 🕖 07:41
Presenter: Jim Marino

Jim’s Component Walkthrough Summary

🔄 Two-Tier Architecture: Control Plane vs Industrial Data Plane

• Control Plane: Lightweight, policy-driven layer managing catalog, contract negotiation, and enforcement.
• Data Plane: Handles big-data, file transfers, and streaming workloads.
• Can be split across clouds or run side-by-side, with DSP messages keeping both planes in sync.

🧱 Industrial Data-Plane Blueprint

• Reference stack combines:
  • On-prem ingest pipeline (OPC UA, AAS, edge filtering, blob/SQL storage)
  • Multi-tenant SaaS layer exposing assets over REST or AAS APIs
• Designed for connector-based transfer

🔐 DSP + DCP Protocols

• Dataspace Protocol (DSP): Orchestrates catalog → contract → transfer
• Decentralized Claims Protocol (DCP): Adds verifiable credentials so only trusted partners get access — no central identity provider required

🧪 Virtualised EDC Runtime (EDC-V)

• A virtualisation container enables:
  • Many tenants + many dataspaces
  • One EDC instance per environment
• Context loaded per request, with message queues ensuring scalability and elasticity

🚀 Efficiency Boost via State Machines

• Swaps inefficient polling (“tick-over”) for reliable message queues
• Enables horizontally-scalable connector functions and reduces CPU waste
• JVM overhead amortised across tenants

🧵 Connector Fabric Manager (CFM)

• Tenant & Ops Manager:
  • Tracks tenants, memberships, DIDs
  • Declarative “deployment definitions” for provisioning clusters
• Pluggable agents deploy or migrate clusters (e.g. Kubernetes cells)

⚙️ Built-in Workload Automation

• Tasks like startup, scale, failover, migration simplified to:
  • Copy config + route traffic
  • Infrastructure handled outside EDC using CFM provider plugins

🛡 Policy Engine Upgrade

• Dynamic CEL-based evaluator supports runtime rules per dataspace
  • e.g., apply different access rules for Catena-X vs. Manufacturing-X
• Enables multi-tenant, policy-isolated runtimes

🏛 Standards & Governance Alignment

• Built to align with:
  • ISO 20151, CEN JTC 25 Trusted Data Transactions
  • Eclipse Dataspace Working Group specs
• Supports procurement-ready governance and future-proof interoperability

🧪 Next Steps for Partners

• Help dogfood the virtualised build
• Contribute data-plane plugins
• Integrate CFM with your cloud
• Public Minimum Viable Dataspace repo and demo clusters are available

🔗 Demo Repository:
github.com/paullatzelsperger/MinimumViableDataspace/tree/feat/aas_demo