VeriFed: Temporally Consistent Continuous Cross-Chain Data Federation

Cross-chain analytics increasingly demand continuous joins across ledgers with asynchronous state evolution. Existing solutions, however, typically assume static snapshots or neglect temporal alignment, yielding semantically inconsistent results when epochs drift. This paper introduces VeriFed, a system for temporally consistent continuous cross-chain joins. We formalize the problem of snapshot-aligned continuous joins, design a Unified Adapter Layer (UAL) to align finalized snapshots across heterogeneous protocols, and develop incremental verification that composes per-chain proofs into a global summary via the Epoch Attestation Mesh (EAM) and the Delta-Linked Proof Forest (DLPF). To sustain high-throughput execution, VeriFed further adopts an incremental multi-objective optimizer that balances latency and monetary cost. Experiments on Ethereum transaction data with a simulated wide-area network (WAN) demonstrate that VeriFed achieves sub-second per-epoch latency (approx. 38 ms) and reduces verification overhead by orders of magnitude compared to state-of-the-art baselines, while effectively detecting tampering with zero false positives. These results confirm consistent efficiency and verifiability under continuous updates.