Dave Syer - Pivotal
In this presentation we take stock of some of the myths and assumptions regarding the runtime costs of using Spring. The conclusion is that preconceptions about slowness, component scanning, autoconfiguration, binding, use of reflection are all completely wrong. We look at data from a large set of benchmarks, analysing the performance of Spring applications. We concentrate on JVM startup time, but data are also available regarding total resource usage, and other runtime overheads. Spring is not slow, despite many people’s preconceptions, but measuring the performance can bring useful optimizations, some of which we will point out in this session. Finally, we can go deeper than that and show what kinds of apps are slower to start and why.