Eclipse GlassFish: This Isn't Your Father's GlassFish
foojay.io
foojay.io
Many people rely on vibes from the past instead of updating their knowledge with the current info. It's true that some companies in the past attempted to taint GlassFish to promote their alternative products. And there was nobody to defend GlassFish and keep it up to date.
This is different now, with GlassFish at the Eclipse Foundation, the OmniFish company behind it and providing enterprise guarantees, and GlassFish itself modernizing with fast startup, runnable JAR, support for latest Java and Jakarta EE, Jakarta Data and NoSQL databases.
Which are already old enough to drive… latest my ass.
I never thought of Spring (or Spring Boot) as good technology, but even for the right audience Spring boot is as exciting as React is exciting for frontend developers or C is exciting for kernel developers. It's a pretty established technology that was new and cool at one point and has just become commonplace and boring for better or worse (I would argue worse, but that's just me).
IMO the kind of person who only knows Spring and doesn't understand modern JEE is exactly the kind of person you don't want to recruit.
The same is true for Micronaught or Quarkus. Learn the frameworks. But they are not a new open standard.
People don't really talk about Jakarta EE as "the standard". Haven't been doing that for quite some time.
You learn it so you don't hand Spring the ultimate monopoly. I thought we all didn't like monopolies? Why give Broadcom one?
Newer frameworks like Quarkus are specifically built for container usage and applications built with it are a bit faster and smaller than Spring boot.
It looks like the industry is moving away from architectures like EE. The desire now is more like a Go deployment: single, self contained deployable. There are make frameworks that support this goal. Maybe EE is one of them, but it's not essential like it was in the early 2000s.
EE doesn't exclude that model at all.
For the most part, maybe like 99%, the Jakarta APIs are agnostic of what the deployment model is. They are APIs to validate input, service HTTP requests, store data in databases, look up roles, connect to identity providers, etc etc.
FWIW Wasm is hitting kubernetes because that's what customers are explicitly asking for, and the majority of enterprise Wasm-on-k8s afopters are doing so precisely because they want to eradicate Spring bloat and the associated supply chain risks from their engineering orgs.
Also, are you sure you are talking about Kubernetes performance over there?
But is it what everybody really wants? To have a single choice? GlassFish provides a choice for those that don't want to become stuck with the "only" option that everybody uses. Java itself provides a lot of options - Oracle JDK, Azul JDK, Corretto JDK and many others. And that's a good thing. Options in frameworks and application servers are a good thing too. The best option wins. Except, there's rarely the best option for every case. And it's good to have all the other options too, in case the most popular option isn't a good one for you.
Spring just made things easy. It also had only one implementation, so you couldn’t be confused because a deployment descriptor worked on one platform and not another. Some random blog post on setup always worked when Spring was in use. If you were following a WebLogic cookbook, good luck on Websphere.
In the end, Easy always wins. Make something hard easier to use, the world will beat a path to your doorway.
Back when JEE was still proto-matter, we had the early versions of JBoss, but it was very raw, and hard to use.
Sun released the Sun Java Application Server 8 (no idea where 1-7 were). It was closed source, but free license for production. It was a much nicer out of the box experience with the platform.
They then rewrote that into Glassfish 1. Free to use, open source, great web UI, great CLI. It was both the JEE reference platform and designed for production.
GF evolved over the years, keeping up with the JEE standards. They refactored it on top of OSGI to improve start up and modularity. If you were so motivated, you could do hybrid OSGI and JEE apps on top of GF.
I used GF in production for years. Never had any big issues with it. The JEE standards worked for us when we ported a large, legacy Weblogic app built in early 2000s to GF, perhaps, 10 years later. All of that legacy sharp pointy XML came right over to the new server, it still supported the original assemblies.
GF suffered from the JEE exodus out of Oracle. Oracle has been an absolutely amazing steward of Java, and I'm not even suggesting that they shouldn't have parted out JEE like they did. But that was a rough patch, and GF withered. Payara pretty much picked up the standard and ran with it from the GF code base, and they've been very good for it.
Its nice that someone is working to keep GF production. Originally, GF had a pretty nice clustering facility that has since been removed. I think it's more an attitude to focus on a production oriented system rather than a reference system that's important, more so than, perhaps, higher end features. Just focus on stability and performance.
I use jersey+glassfish to build very small micro-profile applications. It's stable, small and works.
Not a fan of the HK2 dpendency injector though. Maybe that's my general dislike of how convoluted the spec and implementation (of EE di) is.
I hate how sprawling the (other) implementations are, no it is not ok to pull in 90mb dependencies to support things I don't need. These app servers tend to grow into huge uncontrollable messes. Nobody uses standalone containers anymore and forcing people to pull in all or nothing for the embedded version is just asinine engineering.
* How many applications actually target multiple JEE servers
* Whether stuff like Glassfish and JBoss have to spend as much time selling the paradigm as the product
Personally speaking neither company I've worked out used JEE. We used Tomcat at the last place and the Play Framework at my current place.
I'm not sure that the benefits of long-running Java application servers that you can load and unload JAR-applications from exist (especially when unloading JARs has always been a mess).
Running an app server for a long time and redeploy apps to it is just one of them. To be honest, rarely used nowadays.
Many Jakarta EE products support:
* deploying apps on startup, just like Tomcat * bundling the server into a self-contained app, just like what SpringBoot does with Tomcat * running an app from command line, which Tomcat doesn't support - you have to drop the app into a predefined directory, SpringBoot doesn't support it either - the only option is to bundle the app and the server together during build
Some app servers are very lean, start in seconds, just like SpringBoot. Yeah, Tomcat starts faster, but only with a small app. If you add more libraries, you'd likely get to the same startup speed as SpringBoot or Embedded GlassFish.
I think the perception that JEE means big app servers where you deploy multiple apps and rarely restart the server, is very outdated. Nobody really uses Jakarta EE like that anymore. In fact, Jakarta EE is just APIs, the implementation can vary. Quarkus and Embedded GlassFish are perfect examples. Quarkus, although not fully Jakarta EE, can even build apps to native binaries. And Embedded GlassFish can run the same apps designed to run on app servers, on command line, without any installation of an app server.
Jakarta EE, even with its latest updates, comes from a different world. Standardized library API with interchangeable implementations that are injected by the application server. But wait! We flip the script by embedding the application server inside a fat JAR and shipping everything in a single docker/OCI container. A lot of the stuff that used to happen in the application server (load balancing, shared connection pooling, configuration, service discovery, service bus) happens now at the cloud infrastructure level.
You can still use a MicroProfile-based framework, and Quarkus (which is based on MicroProfile) is very popular nowadays, but once you went along with a certain framework, you're not very likely to replace it. Standardization was the selling point of Java EE in the past, but in the microservice world when you're only betting a smallish microservice on Framework X, people are not so concerned about putting all their eggs in one basket anymore.
However, the problem with that is that it requires writing everything in Java - heterogeneity breaks the model. A language-agnostic solution like containers was bound to win out, it’s just that nothing remotely close existed at the time.
Keep in mind a lot of this was developed even before VMs were commonplace. The first true, usable VMs for x86 were released in 1999, four years after Java’s debut.
(i.e. in the same space as Jboss/Wildfly, WebSphere, etc)
Historically, it was also the reference implementation application server for J2EE.
Tomcat itself has never been an official reference implementation of anything.
Tomcat implements various Jakarta EE APIs, most centrallly Servlet indeed, but also JSP (Jakarta Pages) and JSTL (Jakarta Tags), WebSockets and Jakarta Authentication.
The initial Tomcat was donated to Apache, then used back in GlassFish. GlassFish WAS the reference implementation of Servlet (among others), so indirectly Tomcat kinda was the reference implementation indeed. But just a fork of its code via GlassFish.
It allows running and manage applications on a server, which provides resources to the applications. And it also allows building standalone Java applications, with the server embedded in it, in a way that you would expect from a framework.
On top of that, it provides standard Jakarta EE APIs, so your applications don't need GlassFish, you can run them on other servers too. Or you can easily migrte from other servers and frameworks to GlassFish if you like it more. And you can learn Jakarta EE APIs even before you will use GlassFish, or hire somebody who already knows it even though they never used GlassFish.
It solves some of the same problems you might reach for Kubernetes or OpenShift for, your application gets access to external resources in structured ways and you get to look at dashboards.
GlassFish is an example of such an application server. WildFly is more common, and is the artist formerly known as JBoss. If you have some knowledge in the enterprise Java ecosystem you can quickly and easily (or maybe not, it depends) deploy your creations into these.
In Biodiversity, a glass fish includes a few group of Asian fishes that show crystal transparent bodies to hide from predators. Specially when young. They are vertebrates that evolved transparent muscles. Two gens are kept in aquariums: Parambassis ranga and several ghost catfish from gen Kriptopterus.
We can assume that the programmer likes aquariums. The word Yakarta is not random, as is related with the catfishes distribution.
https://www.theregister.com/2018/03/04/java_ee_is_now_jakart...
if curious (or fomo) it would have taken you about 15 secs to find out what glassfish is, which is still probably 15 less than what you wasted on this mini rant. from there it's up to you to go down the rabbit hole.
However, at a brief glance back at the article, The second sentence in the first paragraph says it's an "application server". Further below the illustration image, there's a text in bold that says "Eclipse GlassFish is now a production-ready, enterprise-grade platform".
So I'm really curious, whether the article didn't make it clear, or there was a lack of interest on your side.