Generated by GPT-5-mini| OpenSplice | |
|---|---|
| Name | OpenSplice |
| Developer | ADLINK Technology (formerly PrismTech) |
| Released | 2002 |
| Latest release | (varies by edition) |
| Programming language | C, C++ |
| Operating system | Linux, Windows, VxWorks, RTOS variants |
| Platform | x86, ARM, PowerPC |
| Genre | Middleware, Data Distribution Service |
| License | Commercial, Open-source (some editions) |
OpenSplice is a commercial and open-source implementation of the Data Distribution Service (DDS) standard for real-time, distributed systems. It provides low-latency, high-throughput data connectivity used in domains that require deterministic behavior and scalability. The software is integrated into ecosystems for embedded systems, aerospace, automotive, telecommunications, and industrial automation.
OpenSplice implements the Object Management Group Data Distribution Service (DDS) specification to enable publish–subscribe communication among distributed applications. It interoperates with standards and technologies such as CORBA-based middleware, Real-Time CORBA, and networking stacks used by SpaceX-class systems and NASA missions. Vendors and integrators pair it with operating systems like Linux distributions, Windows NT, and real-time kernels such as VxWorks and QNX for use in projects from Airbus avionics to Siemens industrial control.
Development of OpenSplice began in the early 2000s by PrismTech, a middleware company later acquired by Avatar, then by ADLINK Technology. Its evolution tracks the maturation of the DDS standard by the Object Management Group and adoption in initiatives such as OMG DDS-XTYPES and AUTOSAR. OpenSplice has been deployed in high-profile programs that reference standards from RTCA and EUROCAE for avionics, and has been cited in academic work from institutions like MIT, Stanford University, and Carnegie Mellon University. Commercial partnerships and acquisitions connected it to corporations such as Thales, BAE Systems, General Dynamics, and Honeywell.
OpenSplice's architecture centers on a distributed set of daemons and libraries that implement DDS entities: DomainParticipant, Publisher, Subscriber, DataWriter, and DataReader. Key components include the core persistence engine, transport adapters, and quality-of-service managers which integrate with real-time operating systems like VxWorks and networking frameworks used by Cisco Systems and Juniper Networks. The runtime supports language bindings in C++, Java, and can be paired with serialization schemes used by Google's Protocol Buffers or Apache Thrift. Integration points include tools for model-based design from MathWorks and systems engineering flows aligned with INCOSE practices.
OpenSplice provides deterministic delivery through configurable Quality of Service (QoS) policies such as reliability, durability, and deadline which align with requirements in DO-178C and ISO 26262 projects. It supports content filtering, automatic discovery via multicast and unicast compatible with IETF networking standards, and persistence mechanisms for fault tolerance used in telecommunications and financial services platforms. Management and monitoring integrate with operations frameworks from Red Hat and Microsoft Azure Stack use cases, while security features map to profiles influenced by OMG DDS Security and IETF TLS.
OpenSplice is used across sectors requiring real-time data distribution: avionics suites in platforms from Boeing and Lockheed Martin; autonomous vehicle stacks developed by firms like Waymo and Aurora Innovation; industrial automation systems by Siemens and Schneider Electric; and satellite ground systems operated by agencies such as ESA and JAXA. It supports robotics frameworks influenced by ROS and research projects at laboratories like Lawrence Livermore National Laboratory and Los Alamos National Laboratory, and is incorporated into defense systems alongside vendors such as Raytheon and Northrop Grumman.
OpenSplice has been offered in multiple editions, including commercial enterprise editions by ADLINK Technology and community or academic editions distributed under open-source licenses in earlier releases. Licensing models have ranged from proprietary commercial subscriptions to permissive open-source distributions that enabled academic adoption at universities including UC Berkeley, University of Cambridge, and ETH Zurich. Corporate deployments often combine paid support contracts with certification processes referencing standards from IEEE and ISO.
Performance tuning for OpenSplice focuses on throughput, latency, and determinism, and benchmarking exercises often compare it with other DDS implementations from vendors like RTI and open projects influenced by Eclipse initiatives. Benchmarks use reference workloads similar to those in SPEC suites and real-time benchmarks used in LINPACK-style evaluations, targeting sub-millisecond latencies for intra-data-center links and sustained multi-gigabit throughput over accelerated network interfaces from Intel and Mellanox. Results are published by corporate engineering teams and independent labs at conferences such as Embedded World and IEEE Real-Time Systems Symposium.
Category:Middleware