A Deep Dive into TENSOR’s Technological Innovations: Reflecting on the software architecture choices: Why blockchain?

Author: Thridium

Moving closer to the final stretch of the TENSOR project’s lifecycle, it is an excellent opportunity to reflect back on the choices made and the technologies used as part of this project. In this brief series, we will be discussing and showcasing some of parts of the TENSOR platform we consider more innovative or used in original ways in the fields of biometric analysis and international collaboration. 

What better way to kickoff this series then, than with one of the more innovative aspects of the project, secure biometric data exchange among different LEAs across Europe, and more specifically the blockchain technology leveraged to ensure the trust and accountability among different actors in the whole process. 

From the very beginning, it was clear that we needed a decentralized, peer-to-peer solution for the data exchange and the related functionality, which is quite a far cry from the standard systems and procedures used typically by European LEAs that follow a more centralized structure, with checks and balances being enforced by each country in a separate manner. Instead, in the pan-European scene, the LEAs operate mostly as equals, leveraging bidirectional support and communication to assist each other in their efforts and ultimately produce a result that is more than the sum of their separate abilities. 

The key factors in such a communication are timeliness and trust. The information exchanged between any two different organizations in critical scenarios needs to arrive in a timely manner to be acted upon, but also the receiver of the communication needs the trust the sender enough, so they are willing to act based on the information they receive. 

Combining the above requirements of decentralization, equality, timeliness and trust, and combining them with standard requirements in the law enforcement field such as transparency and accountability, paints a picture that fits a blockchain-based solution perfectly.  

• The data itself is not stored centrally in one location, under the aegis of a single agency, but each agency has a part of the distributed ledger in their own infrastructure called a blockchain node, thus ensuring decentralization for both governance and network resilience purposes. 

• Any disagreements on the state of the ledger itself are resolved by reaching a consensus among the nodes themselves as equals, and not a central authority presiding over it. 

• The immutability of the ledger itself (guaranteed by both the strong cryptographical connection between blocks and the need for consensus to make any changes) promotes both trust and transparency, as all the users can be sure that the information they see was not tampered in any way. 

• The blockchain also offers a very high degree of auditability, and by extension accountability, ultimately increasing trust, since all new blocks added in the ledger are cryptographically signed by the submitting users, using their privately held encryption keys and timestamped at their time of inclusion in the ledger, essentially leading to the creation of a concrete, traceable and verifiable audit trail. 

• The only requirement that is seemingly antithetical to the usage of a blockchain is the increased speed of the system, since due to the added overhead blockchains are typically very slower compared to databases, but by leveraging smart contracts which offer increased automation, access control and real-time notification mechanism, a lot of this time can be “bought back” by reducing the expected inefficiencies in the data exchange process, thus making a blockchain-based solution adequately performant in real world applications, even though it’s suboptimal on paper.  

Thus, the Data Sharing Platform (DSP) component of the TENSOR platform was born, leveraging the blockchain to greatly enhance the data exchange process. 

In a typical usage scenario of the DSP, the investigators use other distributed analysis components to retrieve possible suspect matches outside their own organization (stay tuned for more on these) and can submit a signed and timestamped profile access request on the blockchain to the data owner, who can choose to respond positively or negatively based on different criteria. If the response is positive, a secure communication channel is opened, and the requestor can access the suspect data for the timeframe allowed by the owner. The biometric data itself is not stored in the blockchain, but rather exchanged privately by the LEAs, but both the request, owner’s response and any access attempts, both successful and prohibited are recorded in the blockchain, and can be cross-checked by both parties or even third-party auditors (such as Europol) to ensure the compliance of both sides to the initial agreement about access rights. 

Hopefully, the inclusion of blockchain technology in the DSP will not only provide a reference template for future streamlined and automated cooperation of law enforcement agencies across Europe, but the lessons learnt will be applied to other sectors in a more general bid to promote open, secure and trustful communication among various stakeholders in many fields.