Research On Optimization Of Caching In Named Data Networking

With the booming of the Internet in recent years,the scale of network traffic is experiencing an explosive growth.The traffic explosion has put a heavy burden on network infrastructure for expansion.How to make full use of the existing infrastructure to cope with the traffic growth efficiently is an important issue.Among all traffic compositions,traffic related to content distribution accounts for more than 90%,the vast majority of which is repetitive video transmission and file distribution.Therefore,the integrated scheduling of content and network is essential for network development.In response to this trend,Named Data Networking(NDN)is proposed,which natively supports the deployment of content name-based forwarding and caching in its architecture,allowing popular content to be cachingd in the local network and user requests to be served directly by locally cachingd content during forwarding,thereby reducing the transmission of redundant traffic in the network.Although caching in NDNs offers the promise of overcoming traffic explosion,traditional NDN still faces the problems of inefficient table lookup and forwarding for cachingd content,less reasonable caching placement decisions,and limited caching scheduling.In this paper,a series of researches have been conducted to address the above problems,and the research results are as follows.1.We designed Adaptive Compression Trie based Bloom Filter(ACTBF),a request filter to quickly filter out uncached content requests for further forwarding.This mechanism reduces the number of meaningless table lookup operations in the caching table for uncached requests,relieves the load pressure on the caching table lookup,and improves the efficiency of NDN table lookup and forwarding based on content names.2.We designed a service-aware optimal caching placement policy Service-Aware Optimal Caching Placement(SAOCP)and a request aggregation-aware placement policy PIT Aggregation Aware Optimal Caching(PAAOC).Considering the mechanism of network layer along-route caching native to NDN that distinguishes it from traditional application layer caching and the feature of caching transparency to forwarding policy,the paper proposes the caching service location-aware caching placement policy SAOCP,by combining the along-route caching mechanism,network forwarding policy,the underlying topology,content popularity,and other factors to make integrated caching placement decisions,which improves the caching hit rate of caching placement and reduces cross-domain traffic generation.In addition,by further incorporating the request aggregation capability provided in the NDN forwarding structure into the caching placement model,the content request rate after network-wide aggregation is derived,whereby a PAAOC caching placement policy with better performance is proposed.3.We proposed an NDN caching system with joint scheduling between network devices and user terminals,a blockchain-and smart contract-based incentive mechanism called Caching and Delivery Market(CDM)for caching resource provisioning,a distributed consensus mechanism called partial Practical Byzantine Fault Tolerance(pPBFT)for content distribution,and a reinforcement learning-based Incentive-Aware Caching Placement(IACP)polciy.This paper proposes an NDN caching system with joint scheduling of both network devices and user terminals to address the problems of caching deployment concentrated in the network infrastructure and lacking collaboration among users,which leads to limited caching scheduling and inefficient caching utilization in the whole network.In order to solve the problem that the caching performance is limited by the willingness of caching providers to share during caching scheduling,this paper proposes a caching resource provisioning incentive mechanism CDM based on blockchain and smart contracts to incentivize the willingness of caching providers.In response to the need for high-frequency verification of content distribution in the incentive mechanism,an efficient content distribution consensus mechanism pPBFT is proposed,which solves the problem of low consensus efficiency and overcomes the difficulty in supporting content distribution consensus.Further,based on the given caching incentive mechanism,the caching placement policy IACP for sharing willingness maximization is proposed to maximize the caching placement performance by maintaining the caching sharing willingness of caching providers at a high level.