Fejltolerante

Fejltolerante, or fault-tolerant, describes the ability of a system to continue operating in the presence of faults. A fault-tolerant system minimizes downtime and data loss by detecting faults and employing alternative resources or degraded modes to maintain service.

Key techniques include redundancy, where critical components are duplicated; failover, the automatic switching to a standby component; replication, where data is copied across multiple nodes to prevent loss; checksums and parity for error detection; and graceful degradation, where the system continues with reduced performance rather than failing completely. Architectural approaches such as N-modular redundancy, hot or cold standby components, and distributed consensus protocols (for example Paxos or Raft) coordinate multiple modules to agree on state.

Applications span hardware, software, and networks. In hardware, RAID configurations and redundant power supplies are common. In software, redundant services, load balancing, and isolation of failures help maintain availability; in storage, replication and erasure coding protect data. In distributed systems, multi-region deployments and quorum-based writes aim to maintain availability even when some nodes fail.

Trade-offs include higher cost, complexity, and potential increases in latency due to coordination. Design choices often balance availability with consistency, as seen in the spectrum between strong and eventual consistency. Key metrics used to assess fault tolerance include availability, mean time to failure, mean time to repair, recovery time objective, and recovery point objective.

Fejltolerante design is central to critical infrastructure and services where uptime and data integrity are essential. See also fault tolerance, redundancy, high availability, RAID, distributed systems.