Cardano database artifacts on IPFS - PoC

## Why

Mithril's snapshot distribution is fully centralized today: the aggregator uploads archives to a single Google Cloud Storage bucket, and every client downloads directly from GCS. This means IOG bears 100% of storage and egress bandwidth costs (~150GB+ per full Cardano DB download per client), the system has a single point of failure tied to one cloud provider, and there is no community participation in hosting despite the signing protocol itself being decentralized.

The Cardano Database v2 architecture is a natural fit for IPFS: it already decomposes the database into individual immutable trio archives that are truly immutable, individually hashed, and verified against a Merkle root signed by Mithril certificates. IPFS stores content by its hash (CID), so the integrity guarantee is already built into the protocol — you don't need to trust the IPFS peer serving the data, only the Mithril certificate chain.

The goal of this PoC is to validate the technical feasibility of the IPFS distribution path end-to-end: can the aggregator publish immutable file archives to IPFS and can a client retrieve and verify them? This will surface any blockers (performance, CID computation overhead, gateway reliability, archive compatibility) before committing to a production implementation.