Data-driven protocol-handling for interoperable networking environment

To utilize interoperability which make objects independent of platform and is realized by distributed object environment such as CORBA (Common Object Request Broker Architecture), the authors have been studying a data-driven implementation of networking environment which applications on Von Neumann processors and protocol handling/conversion on CUE (Coordinating Users' requirements and Engineering constraints) series data-driven processors can cooperate without any runtime overheads. Real-time multi-processing in protocol handling/conversion is necessary on multimedia networking environment to keep time constraints in each media. In protocol handling, however, pseudo multi-processing on Von Neumann processor can't maintain time constraints because of runtime overheads by context switching. Therefore, protocol handling/conversion should be implemented on, CUE series data-driven processor which realize real-time multi-processing without any run-time overheads. Besides, it is necessary to cooperate application objects on Von Neumann processors. This paper described an implementation of networking interface which used CUE series data-driven processors and network LSIs to realize real-time multi-processing and programmability for flexible protocol handling/conversion. To minimize overheads in data transmission, the implementation applies FPGA (Field Programmable Gate-Array) and DPAI (Dual Port Memory) to the data transmission. As a result, we demonstrate that the networking interface has achieved maximum throughput of network and discuss possibility of cooperation among CUE series data-driven. processors and Von Neumann processors.