Now that we've seen some valid points that confirm why JVM is a viable platform for modern software development, let's look at some places where JVM usage is particularly popular:
- Web applications
- Big data analysis
- Internet of Things (IoT)
With its focus on performance, JVM is a very popular choice for web applications. When built correctly, applications can scale really well, if needed across many different servers.
JVM is a well-understood platform, meaning that it is predictable. Plus, it provides many tools to debug and profile problematic applications. Because of its open nature, monitoring of JVM internals is also possible. For web applications that have to serve thousands of users concurrently, this is an important advantage.
JVM already plays a huge role in the cloud. Popular examples of companies that use JVM for core parts of their cloud-based services include Twitter (famously using Scala), Amazon, Spotify, and Netflix. The actual list is much larger.
Big data is a hot topic. When data is regarded too big for traditional databases to be analyzed, one can set up multiple clusters of servers for processing such data. Analyzing data in this context can, for example, refer to searching for something specific, looking for patterns, and calculating statistics.
This data could be obtained from the data collected from web servers (for example, logged visitors' clicks), the output obtained from external sensors at a manufacturer's plant, legacy servers that have been producing log files for many years, and so forth. Data sizes can vary wildly as well, but often, they take up multiple terabytes in total.
Two popular technologies in the big data arena are the following:
- Apache Hadoop (provides storage of data and takes care of data distribution to other servers)
- Apache Spark (uses Hadoop to stream data and makes it possible to analyze incoming data)
Both Hadoop and Spark are for the most part written in Java. While both offer interfaces to a lot of programming languages and platforms, it will not be a surprise that JVM is one among them.
The functional programming paradigm focuses on creating code that would run safely on multiple CPU cores, so languages that are fully specialized in this style, such as Scala or Clojure, are appropriate candidates to be used with either Spark or Hadoop.
Portable devices that feature Internet connectivity are very common these days. Since Java was created with the idea of running on embedded devices from the beginning, JVM is, yet again, at an advantage here.
For memory-constrained systems, Oracle offers the Java ME Embedded platform. It is meant for commercial IoT devices that do not require a standard graphical or console-based user interface.
For devices that can spare more memory, the Java SE Embedded edition is available. The Java SE Embedded version is very close to the Java SE discussed in this book. When running a full Linux environment, it can be used to provide desktop GUIs for full user interaction.
Both Java ME Embedded and Java SE Embedded platforms can access the general-purpose input/output (GPIO) pins on the Raspberry Pi, which means that sensors and other peripherals connected to these ports can be accessed by Java code.