18-year-old student builds Rust-based Redis and Kafka alternatives reporting high throughput

An 18-year-old software engineering student from Pakistan, Yasir, has developed two infrastructure tools from scratch in Rust as alternatives to Redis and Apache Kafka. He argues that Kafka's reliance on the JVM causes garbage collection pauses and large memory use, while in-memory stores can bottleneck on multi-core CPUs due to thread contention and locks. His in-memory store, VelocityDB, uses a thread-per-core, lock-free design and io_uring to bypass kernel overhead, and he reports it reaches over 1.6 million requests per second. His message queue, Velocity-MQ, reportedly handles more than 600,000 operations per second with minimal garbage collection overhead. These performance figures are self-reported in his post and have not been independently verified.