InfoQ·2yHow to Minimize Latency and Cost in Distributed Systems
Distributed systems across multiple availability zones can incur high data transfer costs and performance issues. Zone Aware Routing helps optimize latency and costs by directing traffic within the same availability zone. Implementing this requires tools like Istio and support from distributed databases. It's crucial to handle uneven service distribution to avoid hotspots and ensure system reliability.
InfoQ·2yHow to Make Technical Debt Your Friend
Technical debt (TD) is often viewed as a liability, but it can be leveraged to facilitate learning and avoid over-investment. Using a Minimum Viable Architecture (MVA) approach allows teams to make informed decisions about addressing TD by focusing on what is necessary for the success of a Minimum Viable Product (MVP). Not all perceived TD needs to be resolved, as some may turn out to be non-issues. Understanding how to appropriately manage TD helps teams effectively balance short-term delivery goals with long-term viability.
InfoQ·2yHow We Created a High-Scale Notification System at Duolingo
Duolingo created a high-scale notification system to send millions of notifications within seconds during the Super Bowl. They discussed their transition from a monolithic application to a microservices architecture using AWS ECS, Terraform, Python, and JVM languages. Key tools include Zombie mode for continuous user experience during backend issues, Jeeves for monitoring user feedback, and Freeze Gun to maintain user streaks during incidents. The Superb Owl service demonstrated the ability to send notifications quickly and reliably by leveraging asynchronous processing and AWS services.