At the heart of the Java platform lies the Java Virtual Machine, or JVM. Most programming languages compile source code directly into machine code, suitable for execution on a particular microprocessor architecture. The difference with Java is that it uses bytecode - a special type of machine code.
Java bytecode executes on a special type of microprocessor. Strangely enough, there wasn't a hardware implementation of this microprocessor available when Java was first released. Instead, the processor architecture is emulated by what is known as a "virtual machine". This virtual machine is an emulation of a real Java processor - a machine within a machine (Figure One). The only difference is that the virtual machine isn't running on a CPU - it is being emulated on the CPU of the host machine.
Figure One - JVM emulation run on a physical CPU
The Java Virtual Machine is responsible for interpreting Java bytecode, and translating this into actions or operating system calls. For example, a request to establish a socket connection to a remote machine will involve an operating system call. Different operating systems handle sockets in different ways - but the programmer doesn't need to worry about such details. It is the responsibility of the JVM to handle these translations, so that the operating system and CPU architecture on which Java software is running is completely irrelevant to the developer.
Figure Two - JVM handles translations
The Java Virtual Machine forms part of a large system, the Java Runtime Environment (JRE). Each operating system and CPU architecture requires a different JRE. The JRE comprises a set of base classes, which are an implementation of the base Java API, as well as a JVM. The portability of Java comes from implementations on a variety of CPUs and architectures. Without an available JRE for a given environment, it is impossible to run Java software.
Differences between JVM implementations
Though implementations of Java Virtual Machines are designed to be compatible, no two JVMs are exactly alike. For example, garbage collection algorithms vary between one JVM and another, so it becomes impossible to know exactly when memory will be reclaimed. The thread scheduling algorithms are different between one JVM and another (based in part on the underlying operating system), so that it is impossible to accurately predict when one thread will be executed over another.
Initially, this is a cause for concern from programmers new to the Java language. However, it actually has very little practical bearing on Java development. Such predictions are often dangerous to make, as thread scheduling and memory usage will vary between different hardware environments anyway. The power of Java comes from not being specific about the operating system and CPU architecture - to do so reduces the portability of software.
The Java Virtual Machine provides a platform-independent way of executing code, by abstracting the differences between operating systems and CPU architectures. Java Runtime Environments are available for a wide variety of hardware and software combinations, making Java a very portable language. Programmers can concentrate on writing software, without having to be concerned with how or where it will run. The idea of virtual machines is nothing new, but Java is the most widely used virtual machine used today. Thanks to the JVM, the dream of Write Once-Run Anywhere (WORA) software has become a reality.