Multilayer content services in heterogeneous mobile opportunistic networks

Today we are living in the world where networks are becoming significantly more complex and dynamic while the applications are increasingly diverse as the results from the emerging of smart, ubiquitous devices embedded in our daily activities. Applications and services hosted in these smart mobile edges (e.g., augmented reality, self-driving, and various cognitive applications) may suffer from limited network coverage and localized congestion due to dynamic mobility of users and surge of traffic demand. This thesis is concerned with supporting the development of new distributed technologies which are able to support dynamic peer-to-peer content sharing, decrease high operating costs, and handle intermittent disconnections. This thesis proposes CafRepCache, a multilayer collaborative adaptive and predictive caching framework in mobile opportunistic networks. CafRepCache is built on the principle of predictive analytics and heuristics to answer the fundamental multi-attribute decision problem: selecting which nodes represent the best carriers for given contents and deciding which contents to be cached to bring contents as close as possible to the subscribers in order to minimise the end-to-end latency while trying to balance the trade-offs between maximising content delivery while avoiding network congestion and keeping high cache efficiency. Our framework is a fully-distributed and adaptive algorithm that can capture, analyse and predict in real time the spatial-temporal locality of content requests in the dynamic changing connectivity patterns, dynamic mobility of publishers/subscribers and dynamic resources. We extensively evaluate our work using realistic and pseudo-realistic applications and connectivity traces showing that, independent of the network connectivity/mobility and content traffic patterns, our framework outperforms a number of benchmarking and state-of-the-art caching algorithms. Our results show that CafRepCache is able to choose the best suitable set of contents to be cached in the best suitable set of nodes. CafRepCache brings contents closer to the subscribers, minimize the latency of content discovery and content retrieval, keep high cache efficiency and maximise success ratio of content retrieval while avoiding network congestion in heterogeneous mobile opportunistic networks