Dynamic probe based strategy for deadlock prevention in distributed systems

Deadlock detection and prevention of transaction are very important to concurrency and integrity in database systems. Based on investigating the existing strategies of deadlock prevention in distributed database systems, an improved strategy of deadlock prevention was proposed, which used the Dynamic Probe (DP) technology. The DP method created the probe and sent it to the nodes where the deadlock may exist. After receiving the probe, the nodes compare their own information to that of the probe's. The comparison results showed whether a deadlock had been generated. By this way, the presented approach can be employed to prevent the deadlocks. Analysis shows that the proposed strategy increases the effectiveness of the deadlock prevention and the utilization of system resources.