CAVE: Hybrid Approach for In-Network Content Caching

We present CAVE, a novel hybrid algorithm for caching in content-oriented networks (CON). It relies on popularity-based caching and content chunk indexing. CAVE is a combination of two existing caching algorithms: WAVE which is a popularity-based scheme, and CONIC which is an index-based scheme. In CAVE, routers run WAVE and CONIC, with WAVE having priority over CONIC, while clients use a modified Least Recently Used (LRU) caching scheme to store content when instructed by downstream routers and apply WAVE cache suggestion flags for upstream routers before serving content. We conducted several experiments to compare CAVE against several proposed caching schemes including WAVE and CONIC. Our results showed that CAVE outperforms other algorithms scoring the highest cache hit ratio.