Safety | TigerBeetle Docs

submited by
Style Pass
2024-11-10 11:00:05

TigerBeetle is designed to a higher safety standard than a general-purpose relational database such as MySQL or an in-memory database such as Redis:

TigerBeetle handles and recovers from Latent Sector Errors (e.g. at least 0.031% of SSD disks per year on average, and 1.4% of Enterprise HDD disks per year on average) detects and repairs disk corruption or misdirected I/O where firmware reads/writes the wrong sector (e.g. at least 0.023% of SSD disks per year on average, and 0.466% of Nearline HDD disks per year on average), and detects and repairs data tampering (on a minority of the cluster, as if it were non-Byzantine corruption) with hash-chained checksums.

TigerBeetle exceeds the fsync durability of a single disk and the hardware of a single server because disk firmware can contain bugs and because single server systems fail.

TigerBeetle provides strict serializability, the gold standard of consistency, as a replicated state machine, and as a cluster of TigerBeetle servers (called replicas), for optimal high availability and distributed fault-tolerance.

Leave a Comment