Heterogeneous CCN Cache Allocation Strategy With Collaboration Support
Jianhui Lv,Qing Li,Yong Jiang
DOI: https://doi.org/10.1109/access.2021.3052023
IF: 3.9
2021-01-01
IEEE Access
Abstract:Content-Centric Networking (CCN) is regarded as a promising network architecture which has received the distinct interest from many global research communities thanks to its friendly structure, especially its caching technology has attracted the most widespread attention. However, the current caching strategies have some limitations, such as redundant content copies, low cache utilization rate and unbalanced node load. In order to further optimize cache, this paper investigates the classical cache allocation problem, i.e., distributing the cache capacity across some Content Routers (CRs) under a constrained and fixed total cache budget. At first, both topology information and traffic characteristics as two factors to determine the importance of CR, where the evaluation of network topology depends on degree centrality, betweenness centrality and closeness centrality while that of traffic characteristics depends on node load and interest preference. In particular, given the data redundancy due to the high-dimensional feature, increasing the complexity of computation, t-distributed Stochastic Neighbor Embedding (t-SNE) is used to reduce the dimensions of data. Then, for the CR with cache capacity, its Content Store (CS) is divided into collaborative region and non-collaborative region by a heuristic algorithm to facilitate the optimal performance, and the CS with collaboration needs to interact the cached information with its neighborhoods. Finally, the performance evaluation is driven by the real dataset over GTS-CE topology. The experimental results reveal that the proposed collaboration-supported cache allocation strategy is more efficient than two baselines, i.e., increasing cache hit ratio by 7.79%, increasing cache utilization rate by 11.31%, decreasing routing delay by 49.38% and decreasing load balance by 50.98%.
computer science, information systems,telecommunications,engineering, electrical & electronic