TLDR: A research paper evaluates public, private, and hybrid Distributed Ledger Technologies (DLTs) for storing semantic data (knowledge graphs) in data spaces. It finds that private DLTs offer the best performance and storage efficiency, while public DLTs are costly and less efficient. Hybrid DLTs provide a balanced approach, combining private network efficiency with public auditability, making them suitable for scenarios requiring both privacy and traceability. The choice of DLT depends on specific data space requirements like trust models, performance, and auditability.
Data is the new oil, and in our increasingly interconnected world, sharing it securely and reliably is paramount. This is where “data spaces” come into play – decentralized infrastructures designed to enable sovereign, secure, and trustworthy data exchange among many participants. Imagine a shared digital environment where different organizations can exchange information while maintaining control over their own data. This concept is gaining traction, especially with initiatives like GAIA-X and the International Data Space Association (IDSA) pushing for sectoral data space architectures in areas like manufacturing, mobility, and healthcare.
A core principle of data spaces is “data sovereignty,” meaning data owners retain exclusive self-determination over their digital assets. To truly make data spaces work, however, a critical challenge needs to be addressed: semantic interoperability. This refers to the ability of different systems and applications to understand and use data exchanged between them, even if they were originally designed independently. Semantic web technologies, like knowledge graphs, have been proposed as a solution to achieve this by providing a standardized way to model and link data.
Distributed Ledger Technologies (DLTs), often known as blockchain, are a natural fit for the underlying infrastructure of data spaces due to their decentralized and tamper-evident nature. However, efficiently storing and managing semantic data, particularly knowledge graphs, directly on DLT platforms has been a significant hurdle. Many existing approaches either store only references to data on the DLT, delegating the actual data storage to external services, or face inefficiencies in terms of cost and performance.
Evaluating Semantic Data Storage on DLTs
To bridge this gap, a recent research paper titled “Analysing semantic data storage in Distributed Ledger Technologies for Data Spaces” by Juan Cano-Benito, Andrea Cimmino, Sven Hertling, Heiko Paulheim, and Raúl GarcÃa-Castro, presents a systematic evaluation of how semantic data can be stored across different types of DLTs. The study specifically looks at public, private, and hybrid DLTs, using a real-world knowledge graph called KBPedia as its experimental basis. The goal was to compare their performance, storage efficiency, resource consumption, and capabilities for updating and querying semantic data.
The researchers selected Ethereum as a representative public DLT, known for its transparency and decentralization but also its transaction costs (gas). For private DLTs, they chose Hyperledger Fabric, which offers greater control, better performance, and more efficient consensus mechanisms due to its permissioned nature. A hybrid DLT approach combined Hyperledger Fabric for private data storage with Ethereum for public auditability, where cryptographic hashes of data batches are recorded on the public chain without revealing the sensitive data itself.
Key Findings: Performance and Efficiency
The study revealed significant differences across the DLT types:
- Storage Efficiency: Private DLTs proved to be the most efficient in terms of disk space usage. They allow for batching large sets of RDF triples into single transactions, avoiding the strict constraints of public DLTs. Public DLTs, especially when using smart contracts to store data, consumed considerably more disk space due to the overhead of state storage, indexing structures, and event logs. Hybrid DLTs mirrored the private DLT’s efficiency for the actual data, with only a marginal increase for the public hash.
- Write and Update Performance: Private DLTs consistently offered the lowest latency for writing and updating semantic data. Their simplified consensus mechanisms and ability to handle larger transaction batches contributed to faster processing. Public DLTs, particularly with smart contracts, incurred much higher latency and computational load due to the execution logic within the Ethereum Virtual Machine (EVM) and associated gas costs. Hybrid DLTs performed similarly to private DLTs, with a slight overhead from interacting with the public ledger.
- Query Performance: When it came to reading and reconstructing the knowledge graph, private DLTs again demonstrated the best performance. This is attributed to their efficient access to key-value state storage. Public DLTs, especially those relying on scanning the full chain history for direct transactions, were the slowest. While smart contracts improved public DLT query times by leveraging indexed events, they still lagged behind private solutions. Hybrid DLTs were slightly slower than private DLTs because they include a verification step on the public DLT, but still benefited from efficient querying on the private side.
- Gas Consumption: Gas costs are unique to public DLTs like Ethereum. Storing data directly in transactions was less costly than using smart contracts, which involve complex EVM operations and state modifications. Private DLTs do not incur gas costs. Hybrid DLTs had a moderate and fixed gas cost for each public event (the hash), regardless of the data batch size, reflecting the cost of maintaining auditability without exposing the data.
Also Read:
- AI Models Streamline Metadata Creation for Data Catalogs
- Unlocking Smarter LLM Answers: The K2RAG Framework
Choosing the Right DLT for Data Spaces
The research concludes that the optimal DLT strategy for semantic data storage in data spaces depends heavily on specific requirements:
- Private DLTs are ideal for scenarios demanding high performance, control, and cost-efficiency, especially in domain-specific or consortium-based data spaces where participants are known and data sovereignty is paramount.
- Public DLTs offer strong decentralization and transparency, making them suitable for open, trustless environments where auditability of every transaction is crucial. However, they come with significant limitations in scalability and cost-efficiency for large volumes of dynamic semantic data.
- Hybrid DLTs provide a balanced trade-off. They combine the operational advantages and privacy of private networks with the external auditability of public ledgers. This makes them particularly well-suited for use cases where traceability and regulatory compliance are priorities, while still maintaining internal efficiency.
This systematic evaluation provides valuable insights for stakeholders designing DLT-based data spaces that leverage knowledge graphs. It highlights that while public DLTs ensure high transparency, private DLTs are more efficient for frequent updates and fast query responses over dynamic semantic data. Hybrid solutions offer a compelling middle ground. For more technical details, you can refer to the full research paper available at arXiv.org.
