Research And Implementation Of Information Transmission Mechanism Based On In-network Computing

In recent years,with the rapid development of a variety of emerging business needs such as industrial internet,telematics,and autonomous driving,the requirements for network transmission latency have become increasingly high.At the same time,with the emergence of massive smart terminal devices,the data traffic carried by the network shows explosive growth,which is increasingly contradictory to the limited network resources,leading to problems such as large delay in information transmission and data calculation,many redundancies in information transmission,and low efficiency in information transmission,To this end,this paper carries out researches on information transmission mechanism based on in-network computing,reduces information transmission delay and improves information transmission efficiency by designing a new information transmission paradigm,proposing computational task offloading algorithms,and constructing a simulation verification system.The main work of this paper is as follows.(1)A new network service paradigm is designed for the information transmission mode,by converting the traditional single "transmission"mode into a "computing+transmission" mode,changing the traditional network to act as a "pipe By converting the traditional single"transmission" mode to a "computation+transmission" mode,the traditional network acts as a "pipe" service,and makes the network more flexible for packet processing by deploying computation functions into the network nodes.Simulation results show that the new paradigm performs much better in terms of average latency and energy consumption compared to other computing paradigms.(2)Aiming at the in-network computing offloading problem in the"computation transport" model,a divisible task offloading mechanism based on dynamic placement genetic algorithm is proposed to dynamically allocate computation tasks according to the changes of network state and resource state.Simulation results show that compared with the direct offloading to the cloud computing center,it can alleviate the pressure of processing data in the cloud computing center;compared with the latest distributed multi-hop computation offloading scheme,it saves 10%of the average latency overhead,and enhances the efficiency of information transmission.(3)We designed and implemented an in-network information transmission simulation system,implemented key mechanisms and algorithms,developed system functional modules and interfaces,built an image transmission business scenario,conducted simulation verification of the in-network information transmission mode,and tested the information transmission efficiency under the new paradigm.The results show that the proposed paradigm can reduce the redundant information in the network and ease the data pressure in the cloud computing center.