2020/10/06 by Savvas Kastanakis, Pavlos Sermpezis, Kastanakis, Savvas +7
Computer Science · #Caching and Content Delivery #FOS: Computer and information sciences #Multimedia (cs.MM) #Networking and Internet Architecture (cs.NI) #Opportunistic and Delay-Tolerant Networks #Recommender Systems and Techniques
paper · pdf · doi:10.48550/arxiv.2010.03183
openalex publication_date 2020/10/06 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Joint caching and recommendation has been recently proposed as a new paradigm\nfor increasing the efficiency of mobile edge caching. Early findings\ndemonstrate significant gains for the network performance. However, previous\nworks evaluated the proposed schemes exclusively on simulation environments.\nHence, it still remains uncertain whether the claimed benefits would change in\nreal settings. In this paper, we propose a methodology that enables to evaluate\njoint network and recommendation schemes in real content services by only using\npublicly available information. We apply our methodology to the YouTube\nservice, and conduct extensive measurements to investigate the potential\nperformance gains. Our results show that significant gains can be achieved in\npractice; e.g., 8 to 10 times increase in the cache hit ratio from cache-aware\nrecommendations. Finally, we build an experimental testbed and conduct\nexperiments with real users; we make available our code and datasets to\nfacilitate further research. To our best knowledge, this is the first realistic\nevaluation (over a real service, with real measurements and user experiments)\nof the joint caching and recommendations paradigm. Our findings provide\nexperimental evidence for the feasibility and benefits of this paradigm,\nvalidate assumptions of previous works, and provide insights that can drive\nfuture research.\n