Thorntail Community Announcement on Quarkus
Thorntail Community Announcement on Quarkus
It has been approximately one year since we announced the Thorntail 4.x proof-of-concept with the goal of improving performance and improved developer experience over what 2.x offered. It has been great to see the interest and excitement around Thorntail 4.x from the community. Not long after 4.x was announced, there was an internal proof-of-concept at Red Hat looking more holistically at Java’s use in Kubernetes along with vastly reducing resource requirements and improving performance. On the 7th of March 2019, Red Hat announced the Quarkus project with impressive resource consumption and performance numbers, and a fantastic development experience utilizing SmallRye to implement the Eclipse MicroProfile specifications.
Naturally, such an announcement will raise questions about what that means for the future of Thorntail. In particular, both Thorntail and Quarkus utilize SmallRye to implement the Eclipse MicroProfile specifications, are great for microservices development, and can generate uber jars.
As there are both 2.x and 4.x development streams for Thorntail, we will address the future of each separately.
For 2.x we have both community and product releases available, but this post will focus on the community releases in detail. However, Red Hat remains committed to its OpenShift Application Runtimes (RHOAR) customers, who will continue to receive support through November, 2020 for Thorntail 2.x within RHOAR. For additional information about Thorntail 2.x in RHOAR, please consult your Red Hat Account Manager.
From a community perspective, we will continue releases in the 2.x series for the next 18 months as we assist in transitioning existing Thorntail 2.x users to either Quarkus or WildFly. For Thorntail 2.x use cases that don’t fit into the direction of Quarkus, we will be working with the WildFly team to bring these features into WildFly for all users. That will include Eclipse MicroProfile APIs, with implementations from SmallRye, and we are tentatively planning a future version of WildFly offering the ability to generate uber jars in a similar manner to Thorntail 2.x, or an alternative server runtime reduction approach.
The upcoming community releases will continue to include new features, in particular from future versions of Eclipse MicroProfile, as well as continuing to fix and resolve existing issues. We’re still working through what that means for the quarterly release cadence, and will update the community when we have further information.
There has been a great deal of interest in the Thorntail 4.x proof-of-concept since its introduction early last year. Much of this interest was due to the improvements over 2.x with respect to usability and performance.
We, the core team, have greatly appreciated all the excitement and contributions we’ve received from the community towards making 4.x even better. We provided a lot of input to the design and implementation of Quarkus that directly came from the ideas within the Thorntail 4.x stream. For us, Quarkus is 4.x and much more. Thus, with the announcement of Quarkus we are ending development of the 4.x stream.
Although we’re sad to see 4.x end, we’re very excited about Quarkus as it incorporates the usability ideas present in 4.x. In addition, it brings fantastic new features like unified imperative and reactive development, compile time boot, live reload, and generating a native executable of an application using Substrate VM, part of the GraalVM project! Quarkus, with its inherent support of Substrate VM, is also a great fit for serverless environments.
If you haven’t looked at the Quarkus codebase yet, you might be wondering what that means for SmallRye being the Red Hat implementation of Eclipse MicroProfile. Have no fear, we’re not dropping SmallRye. Looking at the Quarkus codebase you will see it includes the SmallRye implementations of Eclipse MicroProfile. SmallRye will be critical to defining the programming model for Quarkus.
As with many projects in open source, as developers we’re always looking to make advancements and improvements that will make the lives of our users easier. From the initial beginnings of 2.x utilizing WildFly internally, we took it as inspiration for the 4.x proof-of-concept. 4.x offered evolutionary improvements over 2.x in terms of speed, memory usage, and ease of development. These ideas and goals were then applied to unified imperative and reactive development, compile time boot, live reload, and generating a native executable for a tremendous evolution above 4.x into Quarkus. Quarkus is the latest evolution of our efforts in developing a framework for microservices, and serverless, that began back in 2015 with 2.x.
The core team would like to thank the community for all their support, feedback, and contributions to the project over the years! During the next 18 months we will ensure existing Thorntail users have a smooth transition over to either Quarkus or WildFly. For those switching to Quarkus, we hope you enjoy using it as much as we have.