I was reading Dan McKinley’s excellent blog post about the lack of a reverse gear in your Internet deployment dumptruck, and he makes some great points about limiting the unexpected impact of changes. Deployment systems are an area that I’ve always been adjacent to but never had a chance to dig into deeply, so I wanted to put some thoughts to paper in hopes of clarifying them. It seems like there…
Howdy friends! In this blog post, I show how Kudu, a new random-access datastore, can be made to function as a more flexible queueing system with nearly as high throughput as Kafka. One of the more exciting recent developments in data processing is the rise of high-throughput message queues. Systems like Apache Kafka allow you to share data between different parts of your infrastructure: for…