The Design And Implementation Of Mutiple Cards Adapter Manager In SS7Network

SS7network is one of the main technologies in current communication filed. As the great speed inflation of network size and business, the maintenance and management work has been an indispensable part in communication system. All data to be sent to upper application must go through different kinds of net devices which exist in monitor system. Above all, it’s of great important to design a system which not only can segregate the difference between hardware, but also can offer a public and unified interface to upper network layer.The aim of data transmission system is to provide reliable support to upper application. However, current data transmission mechanism has occurred miserable packets losing when met high speed burst data flow. And what’more, not only the addition of high-speed card will preempts resources with normal card which causes data transmission exception, but also the contradiction between high speed and too much copy progress creates bottlenecks.Oriented to the requirement of real-time in data transmission, merging with the gist idea of dynamic memory allocation, the paper designed a new memory management scheme, which focus on the strategy of dynamic memory allocation according to speed of data flow, the improvement of memory utilization and the decrease of memory copy times. Based on the research on current system, the paper analyzes the shortage of buffer mechanism. For common speed card, it designs two double circular linked lists to manage the spare buffer space and device buffer space. By comparing the speed of input and output of data flow, the system will mount the spare buffer block to device buffer space dynamically, or recycle the unused spare buffer block from device buffer. It can reduce the system fragment as well as overhead caused by temporary allocation. On the other hand, for the optimization of high speed card, the paper separates its buffer from mutual buffer mechanism and takes advantage of Memory Map technology. At the end of this paper, the testing result shows great decline of CPU utilization and packet losing ratio drop.