2026-10-28 –, Naturalis 01
Current PID systems, especially DOI, still depend on a central authority that could disappear, change its policies, or simply go offline. We present a non-blockchain-dependent, truly decentralised Persistent Identifier architecture that eliminates single points of failure without the energy costs or governance complexity of distributed ledger systems. It is reached through a DHT-based, cryptographically verifiable mesh of nodes, realms, and providers that can mint, resolve, and version PIDs across scientific research, industrial automation, and open-data ecosystems. The architecture supports FAIR principles through persistent, machine-actionable resolution paths, while data loss resistance is achieved through mandatory replication across providers and a Git-like sectioned hash tree that the integrity. The presentation describes a realm-based multi-lookup mechanism by which the same PID could be resolved from multiple contexts. We kindly invite all the participants to discuss the ideation of the implementation and the addressed challenges.
Today's persistent identifiers carry a structural contradiction that every librarian, data manager, and researcher eventually confronts: they maintain permanence, yet they depend entirely on a likely centralised registry whose own future nobody can guarantee. The annual fees, governance negotiations, and singular global providers we are all already paying for it in time, budget, and institutional risk.
This structural vulnerability is not distributed equally. Institutions in the Global South often face subscription fees denominated in foreign currency, complex procurement processes, or outright commercial restrictions that limit their ability to mint or maintain identifiers. It creates an inconvenience for well-resourced institutions, becoming an outright barrier to participation for others.
Our presentation describes a novel PID architecture designed to resolve that contradiction at its main intention. The concept doesn’t rely on blockchain infrastructure, cryptocurrency stakeholders, or any form of integrated registry with its governance burdens. Instead, it draws on well-established, freely available data engineering principles already powering the open internet: distributed hash tables (DHTs) for discovery, cryptographic signatures for integrity, Git-like sectioned hash trees for versioned storage, and message-passing protocols. These components combine into a network where anyone can mint, resolve, maintain, and express trust in certain Providers independently, and without a central authority.
The key innovation is a layered system of realms - think of them as organised, rule-governed namespaces that any minting entity (Provider) can define and maintain for their own data. Because the same object can belong to multiple realms at once, a single identifier intentionally can be looked up meaningfully from different scopes (realms) without duplicating records or paying for multilayered references.
The multi-realm architecture borrows a concept from POSIX operating systems: just as Unix-like systems expose devices through multiple virtual directories (by-id, by-name, by-path) pointing to the same physical device, the DPID network allows the same identifier to be resolved through different realms without duplicating records. This enables what we call instrumental realms: resolution paths that reflect different perspectives on the same object, making PIDs more useful and adaptable to diverse workflows.
Trust in the proposed network is rankless and local: each node independently verifies the integrity of identifiers and metadata using publicly verifiable cryptographic methods. This approach, modelled on verification patterns of Git and the redundancy patterns of BitTorrent, ensures that trust remains a local, earned property rather than a globally assigned status.
Resilience against data loss is built in from the start. Every provider in the network is expected to maintain independent copies of the data objects they resolve, and the architecture is specifically designed so that identifiers remain resolvable even when individual Providers go offline or withdraw from the network.
The project is currently at an active conceptual and prototyping stage. The early-stage node prototype implementation is in progress. The immediate roadmap focuses on assembling a small coalition of pilot institutions willing to run compliant nodes, stress-test multi-realm resolution across real collections, and co-author the first open specification draft.
