3.7.6 Secure Multi Party Computation
3.7.6.1 SMPC over blockchain
Secure Multi-Party Computation (SMPC) is a subfield of cryptography which allows multiple parties to compute a function over their inputs while keeping those inputs private. SMPC is aimed at creating methods for parties to jointly compute a function over their inputs while keeping those inputs private.
Despite its broad applications in privacy-preserving data mining, machine learning, and other fields, SMPC is often constrained by the complexity of secure computation and the need for a secure communication channel. These limitations are addressed by integrating SMPC with blockchain technology (Jiang, Zhou, & Feng, 2022). The transparency, decentralization, and immutability of blockchain are beneficial for SMPC, as it can mitigate the trust issue and provide a secure and transparent platform for computation. The combination of blockchain and SMPC can ensure that:
No single party controls the computation process.
Privacy of the participants' data is preserved.
The computation process is transparent and auditable.
The result of computation is immutable.
3.7.6.2 PS3I over Blockchain
Secure Multi-Party Computation (SMPC) is a vast field comprising various methods designed to solve a multitude of problems. For the practical purpose of this discussion, we will focus on Private Set Intersection (PSI), an important problem in SMPC. The premise of PSI is to identify the intersection of two or more sets without revealing any private identifiers.
Let's consider a practical scenario where we have two datasets, one with customer phone numbers from Entity X and the other from Entity Y. The task is to identify the number of common elements (i.e., phone numbers) in these datasets without disclosing the actual phone numbers. This is where PSI proves its worth.
It's important to note that in our context, we're assuming that the private data is stored off-chain or in systems such as the InterPlanetary File System (IPFS). The role of blockchain here is not to host the private data but to facilitate the finding of intersections. These intersections, representing meta-information, can be utilized for working with anonymized data, conducting research, or evaluating the performance of the parties involved. In our blockchain environment, we adapt a specific method known as PS3I (Buddhavarapu, et al., 2021), a variation of PSI, tailored for decentralized systems. The method allows for the computation of private intersections while preserving privacy and leveraging the inherent advantages of blockchain technology.
The Private Set Intersection Cardinality (PS3I) is a protocol in SMPC that allows two parties to compute the cardinality (size) of the intersection of their private sets without revealing any additional information about the sets. Paillier cryptosystem, a probabilistic asymmetric algorithm which supports homomorphic properties, can be used in PS3I for secure computation.
Here's a simplified description of how PS3I with Paillier cryptosystem works:
Key Generation: The party who initiates the computation (say, Party A) generates a pair of Paillier keys - a public key and a private key. The public key is shared with the other party (Party B).
Encryption: Both parties encrypt their private set elements using the public key provided by Party A.
Shuffling and Sharing: Both parties shuffle their encrypted sets to remove any order information and then share the sets with each other.
Intersection Calculation: Each party calculates the intersection of the two sets. They do this by comparing the elements of the two sets - since the sets are encrypted with the same public key, the identical elements will have the same encryption.
Cardinality Calculation and Decryption: Party A calculates the cardinality of the intersection set, encrypts it using the public key, and sends it back to Party B. Then, Party A decrypts the received data using its private key to obtain the final result, i.e., the size of the intersection set.
Result Verification: Party B can verify the result by comparing it with its own calculation.
The following scheme represents an intricate process to privately calculate the intersection of datasets between two parties using a decentralized environment in a secure manner. The fundamental premise revolves around the utilization of Secure Multiparty Computation (SMPC), in this case, the Private Set Intersection (PSI) method. This operation is performed off-chain due to the large scale of the data.
The computation involves a series of steps starting with data registration and encryption, key exchange, set difference calculation, and final data decryption. The result is a calculated set difference that is accessible to the receiving party while ensuring data privacy and security.
This process is essentially a cryptographic dance, orchestrated meticulously to ensure that no participant gets to peek into the private data of the other while still enabling the computation of shared interests. The use of the blockchain in this scheme is essential to ensure data integrity, security and accountability, as every computation, transaction and data access is recorded and can be audited.
3.7.6.3 Use Case: Marketing Attribution
It is useful to present the scheme of the method using the example of a specific Use Case. Marketing attribution in today's data-driven world often requires collaboration between various entities that hold different pieces of the customer journey data. In this case, we have three participants:
Participant A (Advertising Hub): Connected to a DGT blockchain node, holds information about advertisement views (phone number, date/time, number of views, advertisement product).
Participant B (Bank): Also connected to a different DGT blockchain node, holds spending data of its clients (based on phone number, time, place, spending sum).
Participant C (Data Analyst/Market Researcher): The data consumer looking for a private data intersection to create a comprehensive picture of consumer behavior, correlating advertisement views with spending data.
PROBLEM STATEMENT:
To derive meaningful insights, these entities need to find an intersection in their datasets. However, the data owned by the entities are private and sensitive, thus cannot be shared directly due to privacy laws and regulations. Moreover, a centralized solution would entail a privacy risk and trust issues.
SOLUTION:
A blockchain-based solution using Secure Multi-Party Computation (SMPC), specifically Private Set Intersection (PSI), enables secure and private computation of the intersection of datasets owned by A and B. We propose a special transaction family DPS3IXOR, that performs this computation on-chain.
The blockchain ensures transparency, immutability, and decentralization. SMPC ensures that the raw data never leaves its owner, and only the computation result (intersection) is known to the authorized parties, protecting the privacy of all parties involved.
PROCESS:
Table 26 PS3 Over Blockchain Process (Major step)
# | Participant | Action | Description |
1 | A, B | Data Preparation | A and B prepare their datasets and apply a one-way function (hashing) on their data elements (phone numbers) to create a hashed dataset. |
2 | A, B | Transaction Initiation | A and B create a DPS3IXOR transaction with their hashed dataset and submit it to their respective DGT nodes. |
3 | DGT Nodes | Transaction Processing | The DGT nodes receive the transactions, validate them, and perform the PSI calculation on-chain using the DPS3IXOR transaction family and off-chain Calculator. |
4 | DGT Nodes | Result Generation | The DGT nodes generate a new transaction that contains the result of the PSI (hashed intersection of the datasets) and update the blockchain state. |
5 | C | Data Retrieval | C, as an authorized entity, can then retrieve the intersection result from the updated blockchain state for analysis. |
Last updated
