TrustDS: A Policy-First, Privacy-Preserving Framework for Interoperable Marketplace Data Exchange Across Edge and Multi-Cloud Environments

We present TrustDS, a policy-first, privacy-preserving framework for interoperable data exchange across edge and multi-cloud environments and commercial data marketplaces. TrustDS compiles human-readable consent, licensing, and governance policies into an execution directed acyclic graph (DAG) that schedules privacy-enhancing technologies (PETs) such as differential privacy (DP), secure multi-party computation (SMPC), and trusted execution environments (TEEs) under explicit latency, utility, and cost budgets. We formalize policy admissibility, state a safety property for admissible plans, and provide a revocation protocol that enforces dynamic consent revocation within a configured Δt under standard liveness assumptions. A cost-aware planner co-optimizes operator placement across edge and cloud regions to minimize latency while respecting egress constraints and utility targets. We empirically demonstrate TrustDS using marketplace microdata and authoritative public microdata accessed through AWS Data Exchange/AWS Marketplace, Google BigQuery Public Datasets (and Analytics Hub), and Snowflake Marketplace listings, anchored to primary publishers (e.g., NYC TLC, CDC, CFPB, and U.S. Census). Across five representative workloads, TrustDS improves median end-to-end latency by 13–30% versus centralized transfer and governed clean-room baselines while maintaining verifiable policy compliance, with revocation propagation below 120 ms in the proof-of-concept and median audit-ledger lag of 130 ms. We provide a reproducible workload specification, policy library, and evidence schema to support independent verification and repeatable comparisons.