A lightweight, fast distributed lock service that coordinates multiple processes and servers with a single key
When multiple processes or servers simultaneously access the same resource (accounts, orders, inventory, batch jobs…), contention arises. Ticketing solves this problem with a single binary protocol over TCP — clients acquire a named key, traverse the critical section, and then release it. In the meantime, the server queues waiters in FIFO order, and if a client dies, lease expiry serves as a safety net.
⚡
Binary Frame Protocol
Designed from the start with fixed-width binary fields, values are read directly by offset without parsing — per-request overhead is minimal.
🔑
Fencing Tokens
Each acquisition issues a monotonically increasing token, protecting safely even when the lock server is not perfectly consistent at every moment.
🔄
Non-Stop Cluster
Priority-based active-standby failover. With at least 2 servers, service remains uninterrupted even during restarts and failures.
🌐
Official Clients in 8 Languages
On the same byte-level protocol, providing natural idiomatic APIs for each language (async or blocking).