Async, the new normal?
We live in a world where synchronous is considered the normal way of thinking. After all, most of the languages and libraries are based on synchronous behavior. Almost every time when we design the architecture of a new system, we think synchronous. It’s common to have client applications experiencing delays while waiting for the server to answer. When we think about a large throughput we tend to add more connections to the server in order to solve the problem. But must it always be like this?
This session will try to reveal the basic principles behind sync and async. What are callbacks? How and when can async be faster? What about congestion? Does TCP have a protection mechanism against it? Can you avoid it by design? I will try to shed some light on these questions.