Optimizing Amazon Simple Queue Service (SQS) for speed and scale

This article discusses recent improvements made to Amazon Simple Queue Service (SQS) to enhance its performance, increase capacity, mitigate scalability issues, and reduce power consumption.

Specifically, the article covers:

• An overview of SQS's architecture, including the customer front-end and storage back-end microservices
• The development of a new proprietary binary framing protocol between the front-end and back-end, which replaced the previous connection-per-request model
• The benefits of the new protocol, including reduced latency, increased fleet capacity, elimination of a scalability ceiling, and improved resource efficiency
• Numerical improvements, such as an 11% average reduction in dataplane latency, a 17.4% reduction in P90 latency, and a 17.8% increase in requests handled by the existing fleet
• Conclusion: The article provides insights into AWS's continuous efforts to optimize its services and promises more such stories in the future