In today’s world, data is everything: the lifeblood of industry, communication, commerce, entertainment, and even national defense. But the amount of data in the world is rapidly expanding at an overwhelming pace—according to Statista, some 74 zettabytes of data were created in 2021 alone, up from 59 the year before—making secure storage and recovery a paramount concern for large organizations. If not properly managed, data loss can be incredibly disruptive; for critical scenarios, the results can be disastrous.

Solutions based on blockchain are a promising answer for many cloud storage applications, with the primary advantage of high security. In this arrangement, data is split up into encrypted segments that are distributed across the network and interlinked through a hashing function. However, blockchain-based storage has considerable reliability issues—according to Baidu Company, some 100-200 nodes fail in each storage cluster every day, which translates to a failure rate of 1-2%. When data loss occurs in a blockchain, the entire distributed cloud storage system suffers. Further, maintaining the integrity of the data in blockchains requires significant computational resources.

To address these shortcomings, a team of IEEE researchers has developed a secure data storage and recovery scheme for blockchain-based networks that promises to not only enhance the ability to quickly repair data but also reduce the resource computational overhead of the storage process as well.

“Due to the particularity of blockchain-based industrial networks, data storage management faces enormous challenges,” said Kuan-Ching Li, IEEE senior member, and Wei Liang, IEEE member, from Hunan University of Science and Technology. “This storage and recovery scheme supports dynamicity, fast repair, and update of distributed data with high precision, repairability and reduced resource overhead, enhancing the way data is stored and accessed in Industrial Network 4.0.”

Blockchain-based fault-tolerant distributed storage structure.

The team’s method employs a local regenerative code technology to repair and store data between failed nodes while preserving the data’s privacy. Multiple local repair groups constructed by vector code can fix multiple storage nodes simultaneously, quickly restoring the data to working order in a computationally efficient manner. The method also evaluates whether the encrypted files in the storage system have been modified maliciously, providing additional insight into the nature of the data loss event.

The paper provides an in-depth foundation of related work on industrial blockchain-based cloud storage schemes, highlighting key developments in regenerative code upon which the authors have improved. The authors then delve into the mathematical models and structures for fault-tolerant distributed storage, as well as algorithms for distributed storage and recovery, then propose a novel scheme for data authentication.

Repair process of multiple failure nodes.

Finally, the researchers showcase the results of their experiments, which involve a 51% attack on an industrial network. According to the team, their proposed method shows a marked enhancement in the real-time performance and security of data storage and repair. The experiments yielded a 9% improvement in the repair rate of multinode data and an 8% increase in data storage rate. Overall, the authors believe that their scheme could be highly beneficial to industrial data storage, reducing the repair overhead of local code in data storage while maintaining security and integrity.