On the integration of Blockchain and legacy systems: a framework for high consistency and low latency of distributed transactions

Transactional systems have been designed to deal with specific database models, such as centralized versus decentralized databases like blockchain. This heterogeneity brings many constraints on the integration of the different systems involved in collaborative business processes, leading to sacrificing either consistency or low latency. Hence, a unified approach that can handle high-volume streaming transactions while ensuring data consistency and low latency is required. To achieve this objective, in this work, we analyze the limitations of these existing transactional models, particularly those relying on log-based mechanisms for durability and recovery. Then, we propose a Transactional Framework that unifies diverse transactional models through decentralized, event-driven, end to end transactional mechanisms. Our solution integrates adaptive compensation strategies to correct errors in real time and prevent duplicate processing while efficiently managing streaming data. By consolidating different transaction models into a single cohesive system, our framework addresses also scalability in order to maintain data coherence under high loads. Our evaluation demonstrates that this approach provides a robust, scalable, and efficient solution for modern transactional environments.