Unlocking Efficiency: Sharded Synchronization in High-Frequency Write-Heavy Systems

In multi-threaded environments with frequent write operations, traditional synchronization methods can face performance bottlenecks, limiting scalability and efficiency.

This talk focuses on sharding — a synchronization strategy that allows for more parallel processing by finer-grained locking. The presentation will showcase the practical implementation of sharding on C++, highlighting its strengths and key considerations. It will then dive into comparison data, exploring how sharding stacks up against traditional mutexes, shared mutexes, and some optimized lock-free containers implementations.

Attendees will leave with a practical understanding of how sharding fits into the broader landscape of synchronization techniques, enabling them to make informed decisions for their specific systems.