An autonomous decentralized software structure has been developed to achieve software online expansion and online maintenance and fault tolerance. In this structure, a software subsystem installed in each of the distributed computers has an autonomous operating function for self-management and coordination with the subsystems without extenal direction and/or execution. A data field (DF) concept is introduced to realize autonomy in each software subsystem. All data are broadcast, with attached content code corresponding to its meaning, into the DF without specifying the receiver. Each module in the software subsystem is connected only to the DF and judges whether to receive the data, or not, from the DF on the basis of its attached content code. The module runs independently from all the other modules after receiving all the necessary data from the DF. This autonomous data-driven mechanism and DF structure ensure that an application software module can be independently produced, loaded through the DF, and tested by the data in the DF, and that it can begin execution while the other modules are operating. Every subsystem checks the consistency of the received data. The effectiveness of this software system is demonstrated by its application to real-time control systems for steel production process control.
|ホスト出版物のタイトル||Unknown Host Publication Title|
|編集者||Harold S. Stone|
|Place of Publication||New York, NY, USA|
|出版ステータス||Published - 1986|
ASJC Scopus subject areas