Field Notes on Porting Concurrent C++ Services to Rust

The author shares composite lessons from several C++-to-Rust migrations of concurrent network and ingestion services, blurring identifying details and rounding figures. The central argument is that concurrency, not types or serialization, is the hardest part of such a port, since implicit C++ contracts about threads, locks, and shared state must be made explicit for Rust to compile. The piece notes that codebases often mix std::thread, Boost.ASIO, C++20 coroutines, and hand-written lock-free structures. It recommends a two-to-four-week audit documenting threads, ownership, sharing, and communication before writing any Rust. Both incremental FFI migration and full rewrites are considered, with the choice depending on whether the inherited concurrency architecture should be kept.