HotSpot (virtual machine)

HotSpot (virtual machine)
Name	HotSpot
Developer	Oracle Corporation; originally Sun Microsystems
Released	1999
Written in	C++, Java
Operating system	Cross-platform
License	GNU General Public License with linking exception

HotSpot (virtual machine) HotSpot is a high-performance Java virtual machine originally developed by Sun Microsystems and maintained by Oracle Corporation. It provides runtime services including bytecode execution, Just-In-Time compilation and multiple Garbage Collection algorithms for applications written for the Java Platform, enabling cross-platform deployment on Linux, Windows, macOS and other operating systems. HotSpot is a central component of the OpenJDK project and has been used in enterprise platforms by organizations such as Apache Software Foundation, IBM, Red Hat, Google, and Amazon.

Overview

HotSpot implements the Java Virtual Machine specification to execute Java bytecode, supporting standards maintained by the Java Community Process and overseen by the Oracle Corporation stewardship after the acquisition of Sun Microsystems in 2010. As part of OpenJDK, HotSpot interacts with projects such as GraalVM and tools like javac, jlink, JEP processes and the Java SE platform. It competes and interoperates in ecosystems stewarded by organizations including Eclipse Foundation, Apache Software Foundation, Linux Foundation and Microsoft.

Architecture and Components

HotSpot’s architecture integrates a bytecode interpreter, multiple Just-In-Time compilers, a runtime VM, a classloader subsystem influenced by designs from Java Development Kit predecessors, and a set of native libraries implemented in C++ and assembly language. Key components include the interpreter, the C1 and C2 compilers inspired by research from Sun Labs and academic groups at Stanford University and University of California, Berkeley, the garbage collectors, the class data sharing mechanism used in distributions like OpenJDK builds, and the Serviceability Agent used by vendors such as IBM and Oracle. Integration points tie to build infrastructures like Maven, Gradle, and continuous integration systems used by projects such as Jenkins and Travis CI.

Just-In-Time Compilation and Optimization

HotSpot features adaptive optimization via tiered compilation, combining a fast baseline compiler (C1) and a high-performance optimizing compiler (C2), concepts influenced by research from Sun Labs, DEC Systems Research Center, and university projects at MIT and Carnegie Mellon University. The JIT uses profiling data collected by the interpreter and low-level runtime counters to perform speculative inlining, escape analysis, loop unrolling and other optimizations similar to techniques developed at Stanford University and in papers presented at conferences like PLDI and OOPSLA. HotSpot’s compilation pipeline communicates with tools such as jstat, jmap, jstack, and jcmd for monitoring and diagnostics used by teams at Oracle Corporation, Red Hat, Google and Microsoft.

Garbage Collection

HotSpot provides multiple garbage collectors, including the serial collector, parallel collector, concurrent mark-sweep (CMS), G1 (Garbage-First), and Z Garbage Collector (ZGC), reflecting work from Sun Microsystems research groups and external collaborations with institutions such as University of Tokyo and companies like Azul Systems. GC implementations balance throughput and latency for workloads common to platforms used by Amazon Web Services, Netflix, LinkedIn, and Twitter. The collectors integrate with runtime features such as biased locking and escape analysis and are tuned using options exposed to administrators at enterprises like Goldman Sachs and Morgan Stanley to manage large heap deployments on x86-64 and ARM64 hardware.

Performance and Tuning

Performance tuning of HotSpot involves selecting GC algorithms, adjusting heap sizes, and configuring JIT and tiered compilation flags used by operators at Netflix, Uber, Airbnb and financial firms that run JVM services. HotSpot exposes diagnostic and profiling hooks compatible with JVM Tool Interface and tools such as VisualVM, YourKit, Async-profiler and Flight Recorder (JFR), used by developers at Red Hat, Oracle, Google and IBM to analyze latency and throughput. Benchmarking commonly references suites and events like SPECjbb, DaCapo, JVM Languages Summit and conferences including JavaOne and Oracle Code One where HotSpot performance innovations are showcased.

History and Development

Development began at Sun Microsystems in the late 1990s, influenced by virtual machine research at institutions like University of California, Berkeley, Stanford University and MIT, and commercial work at companies including Digital Equipment Corporation and IBM. After the release of proprietary HotSpot builds, Sun open-sourced major components through the OpenJDK initiative, which continued under Oracle Corporation after 2010 and with contributions from Red Hat, Intel, Google, Azul Systems and others. Milestones include integration of the G1 collector, introduction of tiered compilation, the addition of ZGC, and ongoing collaboration via the OpenJDK Project and proposals managed through JEPs and the JCP.

Implementations and Variants

Multiple distributions and vendors ship HotSpot-based JVMs, including Oracle JDK, OpenJDK builds distributed by Red Hat, AdoptOpenJDK (now part of Adoptium), Amazon Corretto, Azul Zulu, and SAP SapMachine. Alternative implementations and related projects include GraalVM (which integrates a high-performance compiler), the Eclipse OpenJ9 VM from the Eclipse Foundation and vendor-specific forks used by IBM, Oracle, Amazon and Azul Systems. HotSpot’s codebase is tracked and developed in repositories managed by OpenJDK contributors and reviewed through processes involving organizations like Oracle Corporation, Red Hat, and Intel.

Category:Java virtual machines