Low-power Research And Realization Based On NetFPGA Router

Both in academic and industrial field, building a green energy-saving network has become a hot topic. The power consumption of communicatio n networks is quite huge, and the operators spending in electricity cost accounted for a considerable proportion in its operational expenditure. As we all know, Communication networks covered countless routing forwarding base station. In the base station, the typical network equipment for data forwarding is the routing device. The router power consumption takes up a large proportion of the entire network electricity, but in fact, there is a considerable saving space in router power. Therefore, low power consumption research and development for routing equipment has a vital significance to reduce the energy consumption of the whole communication network, and it also become one of key issues in the construction of green network and improving energy utilization efficiency.The average utilization bandwidth of current internet backbone network reaches 30% and the highest average utilization of the bandwidth is less than 45%. However, the current network devices are generally working in full load condition during a whole day to guarantee network traffic to circulate smoothly in unpredictable network environments, which means the designed bandwidth of network devices is much larger than the average use of the network. Our Subject is based on the improved demand in this area, presented the data processing capabilities of network devices can adaptively adjust the frequency and port link control parameters based on dynamic real-time network load, thereby reduce the overall energy consumption of the network.Based on the current status of low-power studies in routing equipment and research of the reference router engineering design in Net FPGA platform, research work of this thesis includes the following three aspects:1) Completed the design and realization of frequency adaptation technology based on traffic statistics, and gave a detailed process design and hardware implementation. Our thesis first proposed the traffic accurate statistical frequency adaptation technique AFCOMBOT, which could fast real-time calculate network device data traffic, combine with the digital frequency divider and adaptive switching mechanism,dynamically adjust the operating frequency of the network equipment. AFCOMBOT implementation was based on the router front-end module, through the accurate calculation of the traffic load, combination with a buffer conversion mechanism and multi- frequency synthesizer submodule, realized the whole design features by Verilog-HDL.2) Completed the design and realization of port- link sleep control based on timing control mechanism. Power down management based on technology timing mechanism, combined with flow sensing technology and time distribution of statistical data, and finally realized low-power management of the router port with Verolog-HDL and Perl programming on Net FPGA platform.3) Based on the above two low-power technology, we tested the routing device performance and power consumption, then comparised and analysed the power data. Performance tests were conducted by forming a loop data stream with the packet generator and the power optimized reference router. Power test was based on the NI-DAQ device for power consumption data acquisition and the Labview software program for data processing. First, functional test was peformed in the optimized router. In ensuring network performance and quality of service provided, further we tested router power, comparisd and analysised the power datas. The test results showed that the improved router can bring considerable energy-saving effect.