The Internet of Things (IoT) revolution has significantly increased the number of sensor-enabled devices in an ever-growing spectrum of intelligent surroundings. A critical component of future IoT scenarios is the necessity for all devices to function cooperatively, openly, immutably, properly, and with performance guarantee. The community has advocated using blockchain databases to address these concerns, however present blockchain designs suffer from performance limitations. The objective of this Ph. D. research is to propose$SCULPT$, a unique permissioned blockchain database system that abstracts machine learning models into fundamental data blocks that are then stored and retrieved from the blockchain. As such, it doesn't store detailed records on a medium, like blockchains, which is fundamentally very slow due to the expensive verification process.$SCULPT$consensus protocol incorporates technological advances, particularly the concept of Proof-of-Federated-Learning (PoFL). The$SCULPT$prototype system is built on the Hyperledger Fabric blockchain platform, which has shown impressive first results.

Nowadays, an increasing number of blockchain architectures provide well-promising protocols for pseudonymous online payments via proposed cryptocurrencies. Most of them suffer from a number of extensibility and scalability issues, as their capacity regarding the number of transactions they are capable of processing per second is limited. Security is also a challenge for this kind of architectures. This paper presents the design and implementation of the Adrestus system, a blockchain-based transaction system with a novel consensus mechanism that is able to tolerate Byzantine faults and is designed to scale without compromising system security. One of the main components of the Adrestus design is a consistent hashing mechanism for the efficient assignment of transactions on parallel regions, called zones, and for solving load balancing problems. We claim that the Adrestus blockchain system scales linearly without compromising system security and achieves its goals without introducing the unnecessary overhead and by eliminating energy and computational waste. Preliminary theoretical simulations and results reflect that Adrestus exceeds the average throughput of the most well-known cryptocurrencies like Bitcoin, and thus, it achieves a higher performance. In this paper, we present this proposed approach along with simulation results and examine the conditions for the proposed fault-tolerant system to meet safety and liveness.

The Internet of Things (IoT) revolution has introduced sensor-rich devices to an ever growing landscape of smart environments. A key component in the IoT scenarios of the future is the requirement to utilize a shared database that allows all participants to operate collaboratively, transparently, immutably, correctly and with performance guarantees. Blockchain databases have been proposed by the community to alleviate these challenges, however existing blockchain architectures suffer from performance issues. In this short paper we propose Triastore, a novel permissioned blockchain database system that carries out machine learning on the edge, abstracts machine learning models into primitive data blocks that are subsequently stored and retrieved from the blockchain. Triastore comprises of two internal routines, namely: (i) Proof of Federated Learning (PoFL), which trains in a distributed manner a global model for the ingested data; and (ii) Blockchain Consensus, which commits this generated model data on permissioned blockchain database. We present a detailed explanation of our data ingestion algorithm with relevant examples and carry out an experimental evaluation with image data from MNIST. The evaluation shows that our proposed data ingestion framework retains high levels of accuracy with low loss in data quality.