conflictdetection

Conflict detection is the process of identifying situations in which two or more entities contend for the same resource or violate a defined constraint, enabling corrective action to preserve correctness, safety, or consistency. It arises in many technical domains, including database systems, distributed computing, version control, scheduling, and collaborative editing.

In databases and transaction processing, conflict detection is central to concurrency control. Optimistic concurrency control detects conflicts when a transaction commits by comparing the sets of data it read and wrote; if two transactions modify the same data, one is typically aborted or retried. Pessimistic locking prevents conflicts by acquiring locks before access. Vector clocks, logical timestamps, and other metadata track causality to detect conflicting updates in distributed data stores.

In version control and collaborative work, conflicts are detected when multiple edits touch the same lines or logical entities. Merge tools attempt automatic resolution; unresolved conflicts require human intervention.

In real-time systems and robotics, conflict detection refers to collision avoidance: systems monitor trajectories and resources to identify when two agents could collide, triggering braking, rerouting, or resource release. In cloud and distributed storage, conflict resolution policies or data structures such as CRDTs (conflict-free replicated data types) help ensure eventual consistency with minimal coordination.

Challenges include handling latency and partial failures, clock skew, scalability to large numbers of agents, selection of appropriate isolation levels, and balancing performance with correctness. The choice of mechanism—optimistic versus pessimistic, eager locking versus late resolution, or deterministic conflict resolution—depends on application requirements for safety, throughput, and user experience.

See also: conflict resolution, consistency models, concurrency control, CRDTs, merge conflict, transactional memory.