An Information Feature Extraction and Rapid Updating Scheme for Knowledge Centric Networking
Qi Zhang, Xiaofeng Jiang, Shuangwu Chen, Jinsen Xie, Jian Yang, Ling Xing · 2019 International Conference on Computing, Networking and Communications (ICNC) · 2019
Knowledge centric networking (KCN) is a new future internet paradigm that directly accesses key content by assigning each piece of the content a unique knowledge identifier. Knowledge transmission is realized via knowledge routing and forwarding. The knowledge-based routing information has more and longer prefixes than the IP-based ones. Therefore, there are more challenges to achieve the knowledge-based forwarding on the knowledge router in terms of shorter latency, low memory consumption, and fast routing table update. In this paper, we propose a tensorial knowledge-based data structure for the knowledge routing table index. The tensorial knowledge-based data structure employs the CANDECOMP/PARAFAC (CP) decomposition algorithm to extract low-dimensional information features, so the massive memory consumption can be effectively reduced. Meanwhile, we propose a distributed update method which divides the large-scale tensor to many sub-tensors during the content forwarding, leading to computing complexity reduction and low search time. Evaluation results indicate that we can restore the data from the low-dimensional information features with no error which illustrates the feasibility of proposed scheme.