Bringing javax.cache'ing to your application

Data / Integration Track
10 years in the making and finally JSR-107 has been finalized. We now have a standardized Caching API for the Java Platform. And that's all ready for you to use in your Spring apps. But what does it really enable for you? What's with these optional features? No transaction support? How do you efficiently make use of this new API to solve real world problems in your application today? We'll look into how get you best started introducing caching into your Spring application to solve real world problems. And, as we explore the javax.cache API in much more details, we'll see how to push the specifications to its limits and... beyond. Whether implicitly through frameworks (like Spring and Hibernate), or explicitly (cache-aside, cache-through, ...). And how to abstract yourself from a given provider when you require more than what the specifications cover (e.g. transactional caches). We'll also cover some implementer specifics you might want to account for when choosing an implementation, especially if you plan to push the spec to the extreme, like when going distributed and caching terabytes of data.

September 16, 2015

16:30 - 18:00
1 hour 30 minutes
When I grow up, I want to be a computer scientist… until then, I write code. Now at @moogmusicinc — Former distributed system & API dude at @terracottatech