TigerBeetle's static allocation and zero-copy I/O eliminate runtime memory overhead
TigerBeetle Core System Architecture: Deconstructing Performance Engineering
TigerBeetle, a financial ledger database written in Zig, achieves sub-millisecond tail latencies by eliminating dynamic memory allocation, bypassing the kernel cache with direct I/O and io_uring, and using a single-threaded execution loop backed by Viewstamped Replication. Its fixed-size 128-byte structs align with cache lines and sector sizes, enabling zero-copy persistence and predictable performance.
TigerBeetle’s core system architecture demonstrates that extreme performance is not achieved by adding complexity, but by systematically removing it.
- g_delgado14
Is it just me or does it not seem like this whole article was llm generated?
The two "sources" are fake links that lead to 404s
- hoppp
I really wish they turned it into a dependency or database framework where users could define their own business logic to swap out the double entry accounting, while reusing all the system architecture and networking features, consensus etc.
Sort of like a new paradigm where opinionated custom databases could be created with arbitrary entry logic built on this stack.
- jorangreef
Joran from TigerBeetle here! I created TB. Happy to answer questions!
- kilroy123
Their simulation is fantastic:
- teabee89
"By [...] utilizing a single-threaded execution loop, TigerBeetle aligns its software architecture perfectly with the physical realities of modern hardware."
Can someone explain why single-threaded execution loop is more aligned with the physical realities of modern hardware ?