The `-ilities` of Software Engineering - How Qualities...
summary
Defines the actual nonfunctional goals for JVM systems and shows how to measure them: throughput and latency as SLAs, capacity planning across 2–96 cores, and why 32‑core concurrency issues appear on an Azure Gen2 VM while Gen3 holds scaling to 96 cores. Compares concrete GC architectures — parallel (throughput) vs ZGC/Shenandoah (low latency) vs G1 default — explains regions, concurrent vs stop‑the‑world work, fallback GC, and recommends using Java Flight Recorder and Universal Scalability Law (Gunther) to predict and debug scaling.
AI-written summary · about
chapters (4)
related talks
45mThe Joy of Building Large Scale Systems
01h 06mHow should we evaluate garbage collectors?
46mCreating Cloud Native Capabilities Using the Latest Features from Java and the JVM
33mFrom Enterprise Architecture to Software Design - Critical Thinking
27mThe Mutability Matrix of Pain
50mWelcome to the Jungle - A safari through the JVM landscape
01h 00mEngineering Software
50mEnergy Efficient Design Principles of Software Engineering