Queues Don't Fix Overload (2014)
Summary
An article explaining why queues are not an effective solution for handling system overload and discussing better approaches.
Similar Articles
Why Queues Don’t Fix Overload (And What To Do Instead)
Explains why unbounded queues are a bug in software systems, using Little's Law and the bathtub analogy to show that queues only absorb variance, not sustained load. Discusses latency death spirals and advocates for backpressure instead.
Job queues are deceptively tricky
A technical blog post exploring the hidden complexity of job queues, discussing why they are deceptively tricky and offering useful lenses for system design such as being wary of queues, limits, and fault models.
We should get rid of average CPU utilization
The article explains why average CPU utilization is a misleading metric for latency-sensitive workloads, using queueing theory and a real-world production incident. It argues for more nuanced monitoring approaches.
Surprising economics of load-balanced systems
A blog post analyzes the M/M/c queueing model and shows that increasing the number of servers in a load-balanced system improves latency at constant per-server load, a beneficial and somewhat counterintuitive result for cloud economics.
OverFlowLight: Real-Time Gridlock Prevention and Traffic Signal Optimization for Urban Intersections
OverFlowLight is a real-time framework that prevents traffic gridlock by detecting queue overflow using multi-modal sensing and inserting dedicated overflow phases via a hybrid rule-based and RL controller. Deployed across 43 intersections, it reduces overflow incidents by 60.4% and increases network throughput by 18.2%.