Pros and Cons of Java Containerization
Everywhere you look today, you see people talking about Docker, Kubernetes, microservices, or CI/CD. But at the heart of those technologies and approaches lies a central concept — containerization. In this blog, we look at the benefits of Java containerization, and offer a comparison of pros and cons for containers vs. application servers for Java applications.
Let's get started with the basics.
What Is Java Containerization?
Java containerization is the process of packaging a Java service or application within a software container, which typically includes everything needed to execute that package.
Containers are often used in the creation of microservices-based applications, and allow for easily-replicable and scalable services within an application.
4 Benefits of Java Containerization
There are a number of benefits for containerized Java, but the biggest four are scalability, flexibility, resilience, and cost. It's important to note that containerization, while beneficial, isn't a one size fits all solution (as we'll cover in the next section).
1. Granular Scalability
Containerization means that both technology, and teams are broken up into smaller sizes, often with one or two teams responsible for production, deployment, scalability issues, performance downtime, etc. From a managerial perspective, that makes it easier to scale up and down teams as needed.
On the software side, that means your application can scale to meet demand as it appears, and scale down once it's no longer needed.
2. Flexibility
Containers also give developers the flexibility to experiment and try new things. That can mean different languages and their accompanying tools, or even the freedom to try new techniques, protocols, or service meshes.
3. Resilience
Resilience should have an asterisk here. If done well, containerization means you can implement service meshes, or resiliency patterns like circuit breakers. That means a lot less exposure to risk for the application at large, since, ideally, a single service can fail without causing a cascading failure for your application.
4. Cost
Because containerized applications can be scaled horizontally, applications (in many cases) no longer need the powerful hardware that enables a high vertical performance ceiling.
Pros and Cons of Containerization vs. Application Servers for Java Applications
For Java, the central question of containerization comes down to whether or not the benefits of containerization are worth the hassle of spending the time and money needed to make that change. After all, with many application servers you get full access to the hardware virtual machine. With these, you also get shared caches and pools because you have access to all the memory.
That means that all the application instances or threads can share the same data. You don't have a lot of copies, so there's less overhead. Then you also have relational databases, which are more familiar for a lot of developers, and, again, you typically get full control over how those databases are used.
| Container | Application Server |
|---|---|
| Granular scalability | Full access to hardware |
| Flexibility to experiment | Shared caches and pools |
| Better resilience | Simple storage |
| Lower cost | Latency |
But one of the big risks for using an application server is that if that app server crashes, your application is crashing. That means that you need to deploy clustering, which is expensive, and carries the risk of bad performance if you have a resource hog within that cluster. You also have to have a vertical hardware ceiling that can accomodate heavy load, but goes underutilized a majority of the time.
| Container | Application Server |
|---|---|
| Tyranny of choice | High impact crashes |
| More moving parts | One bad app can hurt all apps |
| Less familiar for devs | Planning for capacity is difficult |
| Persistence | Language and platform limitations |
Using an application server also typically means you're locked into a language and platform. There's less ability to experiment with other languages, features, technologies, etc. But that also means that it's functionally less complex, since one developer can theoretically work on another area of the application with no language or technology barriers.
Final Thoughts
Making the choice to containerize or develop a containerized Java application isn't as clear cut a decision as it seems. After all, containers, just like microservices, can add complexity to an application and increase the chance at developing a big ball of mud.
Our advice is to weight the pros and cons of containerization carefully, and decide what the right approach is for your team.
Additional Resources
Want to learn more about the history of containers, what benefits they offer for Java applications, and what the future looks like for container technologies like Docker? This webinar from Rod Cope, CTO at Perforce Software, is a must-watch.
Looking for additional resources on containers, container technologies, and microservices? These resources are a great place to start.
- Blog: What Is Containerization?
- Webinar: Making the Move to Microservices for Java Applications
- Blog: Engineering for Microservices Resiliency
- Resource: Exploring Java Microservices
Are redeploy times driving your consideration of transitioning your application to microservices? Be sure to check out JRebel. It allows Java developers to completely skip the redeploy process. meaning developers can spend more time coding, and less time waiting. Want to try it out on your project? Click the link below to grab a free, 10 day trial.