Towards asynchronous conflict resolution through the concurrency patterns abstraction in DAG-based blockchains

Unlike conventional blockchains, which have a sequential model and restrictive participation rules, DAG-based blockchains offer a flexible structure for enabling concurrency. When all nodes participate actively, the concurrency in block generation naturally creates conflicting branches. Corrective approaches often discard valid blocks to solve conflicts, leading to resource wastage. In preventive approaches, transactions are even separated into groups for parallel processing, complicating transaction processing with outcomes reconciliation, workload balancing, and correctness challenges. This last incurs high communication costs and high confirmation delays. To leverage the whole participant’s capacity to process transaction, the asynchrony in participation and concurrency in block creation must be preserved by resolving the conflicts without arbitrarily wasting resources. This paper introduces a new corrective approach in DAG-based blockchain that identifies discrepancies between conflicting blocks. These differences are classified into concurrency patterns which are useful for resolving conflicts among concurrent blocks maintaining the asynchrony of the system and without discarding, or reversing transactions.