Proactive Caching to support mobility in Named Data Networks
MetadataShow full item record
Information-centric Networks (ICNs) offer a promising paradigm for the future Internet to cope with an ever increasing growth in data and shifts in access models. Different architectures of ICNs, including Named Data Networks (NDNs) are designed around content distribution, where data is the core entity in the network instead of hosts. Given the amount of forecast traffic by mobile users, supporting mobility in NDNs to maintain seamless operation is one of the main challenges yet to be resolved. Accordingly, attempts at handling mobility in NDNs in the literature are mostly studied under simplistic or special cases, and relied on content retransmission as a fallback. This is in addition to the lack of benchmarking tools to analyze and compare such schemes. In this thesis, we investigate how predicting the future state of the network can enable seamless support of mobility in NDN. We propose a set of proactive benchmark solutions which exploit location and data traffic prediction to deliver the content of mobile users (both Consumers and Producers) under application delay constraints. Particularly, the network detects roaming users and caches their prospective content ahead of handover events while considering the maximum tolerable delay and network overheads. Unlike existing literature that focused solely on Consumer mobility, we also handle Producer mobility that impacts the content availability and Quality of Service (QoS). Furthermore, we introduce a practical mobility management scheme that is resilient to prediction uncertainties using stochastic optimization. A guided heuristic search algorithm is also developed to provide real-time near-optimal caching decisions instead of commercial solvers that suffer from poor scalability. All benchmark and heuristic schemes proposed in this thesis are evaluated using a comprehensive assessment framework, which is also used to assess the state-of-the-art NDN mobility support. Simulation results show that our proposed solutions maintain user’s QoS during mobility events. We believe that such results drive incentives for deploying proactive mobility management in future NDN.
URI for this recordhttp://hdl.handle.net/1974/22817
Request an alternative formatIf you require this document in an alternate, accessible format, please contact the Queen's Adaptive Technology Centre
The following license files are associated with this item: