users@javaee-spec.java.net

[javaee-spec users] JCache.next (JavaEE alignment)

From: <ondrej.mihalyi_at_gmail.com>
Date: Mon, 26 Sep 2016 15:21:56 +0000 (UTC)

We had the JCache + Java EE BOF at JavaOne, with Hazelcast, Payara and
Tomitribe there, where we discussed what is missing in JCache to align
it with Java EE. Fortunately, Linda Demichiel at JavaOne talk confirmed
that JCache is still a welcom addition to Java EE if the EG can
continue the work to bring it to Java EE. So our effort still seems to
fit the official plans with Java EE.

The discussion suggested that there are three groups of issues that
future versions of JCache could be taargeting:

 - necessary improvements to align with Java EE (we need them to bring
JCache into Java EE)
   - this includes areas such as specifying default JNDI names, define
classloader behavior to isolate cache for separate deployments, cache
descriptor
 - optional improvements (CDI cache scope, asynchronous API, injection
of Cache and CacheManager) - seems that Payara, TomEE and WildFly
already put some effort on their own, and would welcome a standard
approach
 - integration with other specs (JPA 2nd level cache, caching JPA
queries, facelet cache in JSF, JTA to span transaction over both cache
and database, ...) - JCache should provide means to do it, but first
JCache needs to be officially included in Java EE scope

We try to gather issues in this github repo:
https://github.com/Adopt-a-JSR/jcache-javaee/issues
it also contains simple proof of concepts for some issues.

I suggest focusing on the first group of issues to to at least get
JCache into JavaEE in a usable form. At least one of the 3 current spec
leads, Greg Luck, is working on a draft for a maintenance release 1.1,
so once it is public, we can revise if something needs to be added to
be accepted as part of Java EE 8.

The issues in the 2nd and 3rd group would probably target Java EE 9,
but as the plan is to work on Java EE 9 in parallel, we may be working
to resolve them already now.

In the end I want to highlight that JCache ended relatively high in
both latest DZone/JavaEE Guardian and MicroProfile surveys (among first
5 in each survey). There are high expectations from the users and the
community to have JCache in Java EE soon. There are also new subjects
interested in joining the EG
(https://github.com/Adopt-a-JSR/jcache-javaee/issues/3#issuecomment-248
216328).

I hope that JCache EG and their supporters can get additional guidance
from the Java EE expert group and JCP to make including JCache in the
next version of Java EE posible.

Cheers,
Ondrej Mihályi
Java EE developer, lecturer and Payara engineer