Research On Services Cache Placement And Forwarding Strategy Of In-Network Computing In Named Data Networking

In recent years,with the rapid development of the Internet,the surge of data traffic poses a challenge to the current network architecture.At the same time,the focus of users’ behavior on the Internet is gradually shifting to information exchange and content acquisition.In order to cope with the shortcomings of the current network architecture,the content-centric and user-driven Named Data Networking(NDN)emerged.Its features such as ubiquitous caching,flexible routing and forwarding provide great convenience for content acquisition.In order to make full use of the advantages of NDN architecture,it is considered to embed computing functions in NDN to meet users’ demands for emerging computation-intensive services.However,there are still many problems in cache placement and request forwarding of computing services in NDN,such as the low utilization of cache space,the load imbalance of nodes and network congestion.In order to solve these problems,the main research work of this thesis is described as follows:In order to realize the service caching efficiently and reduce the overhead of cache replacement,a cache placement strategy based on the service cache value and the service source is proposed in this thesis.The service cache value is calculated by combining the service popularity and the Combined Recency and Frequency(CRF)value.In order to improve the cache utilization,a service cache value threshold for cache placement is set.Furthermore,according to the source of services and the location of nodes,different cache probabilities are set,which aims to cache services with higher service cache value closer to the user side.Simulation results show that the proposed service cache placement strategy can improve the cache hit ratio and reduce the hop count of the service requests.In order to realize the load balancing and alleviate the network congestion,a service request forwarding strategy based on improved Ant Colony Optimization(ACO)is proposed in this thesis.In order to achieve load balance among computing nodes,the expected heuristic function of ant colony algorithm is updated by the service source evaluation factor which is calculated by the service computing cost and the transmission cost.Furthermore,the pheromone concentration in the improved ACO algorithm is updated by the information of network congestion status,and the forwarding probability of each interface is adjusted to alleviate network congestion.Simulation results show that the proposed service request forwarding strategy can reduce the service response delay,reduce the packet loss ratio and achieve the load balancing.