Research On The In-network Caching Mechanisms And Algorithms In Smart Identifier Network

The current Internet gradually exposed many of its original design’s drawbacks, and these problems cannot be overcome by continuing "patching over patches". Accordingly, in recent years the networking community is actively designing novel Internet architectures from scratch, aiming at fundamentally rectifying one or more of the current Internet’s drawbacks. In this context, the National Engineering Laboratory for Next Generation Internet Technologies in Beijing Jiaotong University has proposed the smart identifier network as a new clean-slate architectural approach which can not only address the drawbacks such as poor security, scalability and mobility that the current Internet faces, but also can greatly improve the utilization of network resources, reduce the network consumption and improve the quality of experience.The smart identifier network leverages ubiquitous in-network caching, regarding it as one of the indispensable roles to effectively reduce the network traffic and access delay, and hence to improve the quality of experience. In this paper, we design new caching mechanisms and algorithms from the perspective of improving the cache hit ratio, controlling the advertisement cost, and increasing the caching gain based on an overview of the existing caching mechanisms. The main contributions and innovations are summarized as follows:1. In this paper, we describe a method of categorizing caching mechanisms. We can classify the existing various caching mechanisms into several categories based on the classification method, focus on their strategies, corresponding optimization approaches and advantages, and simultaneously discuss the remaining research challenges and opportunities. The proposal of this classification method provides the foundation and direction for the future studies of caching mechanisms.2. Aimed at improving the cache hit ratio, we propose a cooperative caching mechanism based on the probabilistic caching. In this mechanism each content router calculates their caching probability in terms of their topological information and uses the caching probability to decide whether to hold a copy of the forwarding content. The numerical analysis and simulation results show that this caching mechanism can improve the cache hit ratio and reduce the cache hit distance, and is easy to deploy in large scale network.3. Aimed at controlling the advertisement cost, we propose a cooperative caching mechanism based on the Min k-center Caching Algorithm and use it to offer the network administrator the freedom to balance the cache hit ratio and advertisement cost, by controlling the number of copies within a domain for a cached content. The simulation results show that this mechanism can significantly improve the cache hit ratio and efficiently control the advertisement cost, achieving an optimal balance between caching efficiency and scalability.4. Aimed at improving the caching gain, we propose a cooperative caching mechanism in which we establish a lookup strategy based on the cache summary table, a replication strategy based on the request path and a replacement strategy based on the caching gain. The simulation results verify that our caching mechanism can improve the cache hit ratio and reduce the false hit ratio, and more importantly, can shorten the average access latency.