Data-Driven Specification Environment for Real-Time Hyper-Distributed Processing

我们已经研究了数据驱动的规范环境，以提高实时超分布式系统提供的实用程序的可重用性。同时，研究了适用于该系统的超集成数据驱动处理器体系结构，并研究了作为超分布式系统之一的TINA（Telecommunications Information Networking Architecture）环境的实现。取得了以下结果。（Data-Driven Specification Environment）实现了面向重用的数据驱动规格说明环境的第一版原型，该环境可以直接从规格说明生成程序。该环境还支持使用符号原型进行副作用检测。该原型已经证明了其有效性的副作用检测和所谓的验证，通过实际的软件开发。此外，为了满足实时处理的性能要求，我们还研究了一种优化算法， 关于我们 程序分配给多处理机的方法。该方法对遗传算法确定的初始任务进行交互式优化。通过实验评价，该方法有效地提高了两者的吞吐量。和周转时间（Super-Integrated Data-Driven Processor Architecture）为了实现超分布式系统的数据驱动，我们在超集成数据驱动处理器上实现了TCP/JP（Transmission Control Protocol/Internet Protocol）协议处理，并在没有任何上下文切换开销的情况下实现了有效的多处理。此外，我们表明，在ATM（异步传输模式）的TCP/IP可以实现超集成的数据驱动处理器原型，这是最小化的片外通信惩罚的超集成和引入的指令调整到TCP/IP处理。（TINA环境的数据驱动实现。TINA-DPE（分布式处理环境）的数据驱动实施。在TINA环境下实现实时多处理。我们已经实现了inter-ORB（Object.请求过程中代理）协议和应用程序接口。我们已经评估了其实时处理的可能性，没有任何运行时控制。少

We have studied the data-driven specification environment to enhance reusability of utilities provided by the real-time hyper-distributed systems. At the same time, we have researched super-integrated data-driven processor architecture suitable for the systems, and studied the implementation of TINA (Telecommunications Information Networking Architecture) environment as one of hyper-distributed systems. The following results have been obtained.(Data-Driven Specification Environment)We have realized the first version prototype of reuse-oriented data-driven specification environment which can directly generate programs from the specifications. The environment also supports side-effects detection using symbolic prototyping. The prototype has demonstrated its effectiveness for both side-effects detection and so-called verification through actual software development. Furthermore, in order to satisfy the performance requirements essential in real-time processing, we studied a optimization m … More ethod of program allocation to multi-processors. In this method, the initial assignment decided by genetic algorithm is interactively optimized. Through the experimental evaluation, the method is effective to improve both throughput. and turn-around time.(Super-Integrated Data-Driven Processor Architecture)For the data-driven implementation of hyper-distributed systems, we have realized TCP / JP (Transmission Control Protocol / Internet Protocol) protocol handling on the super-integrated data-driven processors and demonstrated effective multi-processing without any context switching overheads. Furthermore, we showed that TCP / IP over ATM (Asynchronous Transfer Mode) can be realized by super-integrated data-driven processor prototype which is minimized off-chip communication penalty by super-integration and introduced instructions tuned to TCP/IP handling.(Data-Driven Implementation of TINA Environment.)Data-driven implementation of TINA-DPE (Distributed Processing Environment.) has been studied to realize real-time multi-processing on TINA environment. We have realized inter-ORB (Object. Request. Broker) protocols and a application interfaces for in tem-object communication. We have evaluated its possibility of real-time processing without any runtime controls. Less