LavinMQ: Understanding code to get the best performance
Building LavinMQ in Crystal was an active choice, as maintaining control over the codebase was a high priority. A decade of learning has prioritized fast updates and easy bug fixes. At the same time, choosing Crystal provided the opportunity to enhance application performance.
LavinMQ was built by experienced developers from 84codes, founders of one of the largest cloud messaging platforms, CloudAMQP. By prioritizing the insurance of stable communication and reliable message delivery in and between services, LavinMQ was built to process data quickly and efficiently while allowing decoupled system components to operate independently and scale as needed.
LavinMQ can handle numerous connections between the publisher, the queue, and the consumer, which is often a requirement in e.g., IoT architectures. It’s an open-source message broker, pre-optimized to be extremely fast compared to other brokers, and can easily handle 1 million messages per second.
We highlighted the characteristics that stand out from a technical point of view. How does it manage to handle more than a million messages per second, and have a low memory footprint? Half of it is due to, and we quote:
