Repository navigation
[Proposal] Speeding Up Java Application Startup with AOT Cache #274
Description
Request Issue
No response
Website Section
Learn --> Tutorials --> Speeding Up Java Application Startup with AOT Cache
Proposal Details
Hello!
I have hands-on experience implementing AOT Cache in an open-source Java project, and I would like to turn that experience into an article and share it here.
The article will cover not only the capabilities of AOT Cache but also the practical experience of using it in a real Java project. I plan to explain:
- why consider the use of AOT Cache
- why we chose it instead of AppCDS and how the two technologies differ
- what results we expected
- how we integrated AOT Cache into the application’s build and startup processes
- what limitations and challenges we encountered
- what results we ultimately achieved in practice.
Particular attention will be given to measuring the application’s startup time and RSS memory consumption before and after introducing AOT Cache. The article will also include measurements using AppCDS for comparison. The results will be presented using concrete numbers and charts. We see value in disclosing the benefits of AOT Cache/AppCDS in the real-world project.
I also plan to discuss the trade-offs, including increased memory usage and other associated costs.
I believe this practical case study would complement the existing materials on AOT Cache and be useful to developers considering adopting it in their own projects.
Author References
I've also written articles on Bell-Soft blog: https://bell-sw.com/blog/mastering-reactive-programming-with-spring-data-r2dbc/
And I'm also the author of quite a lot of Baeldung technical articles: https://www.baeldung.com/author/polivakhamikhail
Activity
danthe1st commentedon Aug 14, 2026
I am not sure whether the dev.java site is the right place for case studies as these are often advertise-y and it could set a precedent.
From what I can see in your description, your article would be very focused on how it looked like for you which isn't necessarily something people can extrapolate from.
I generally see there being more value in general (and beginner-friendly) articles on the topics like the existing CDS article. I am not sure whether an additional "use case" article would add that much of value to that.
Disclaimer:
I am just an outside contributor, I am not working at/for Oracle and have no authority here. This is only my personal opinion.
mipo256 commentedon Aug 21, 2026
I get what you're saying @danthe1st, but I think dev.java is indeed the right place for two reasons
The issue with the majority of articles that outline the AOT Cache/App CDS is that they are based on some hypothetical, playground projects that do not really affect real workloads. But in our case we have a real Spring Boot app that is installed and managed in enterprises. And we know the numbers. So I consider that data more valuable.
Axelix Master has an OSS core on GitHub, so I do not really think that there will be much of a problem of advertising, but I hear you, it depends on your PoV. I personally think it is not a big deal at all.
Sure, we can file that article on dzone or something, but I think for features that are solely specific to OpenJDK, I think dev.java is just a natural choice.
That is how I see it :)