Open AccessAdd‐net: adaptive dichotomy based network‐centric cellular to Wi‐Fi offloading
Author(s) -
Venkataraman Hrishikesh,
Dutta Amit,
Joshi Avinash
Publication year - 2017
Publication title -
iet communications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.355
H-Index - 62
eISSN - 1751-8636
pISSN - 1751-8628
DOI - 10.1049/iet-com.2015.0958
Subject(s) - computer science , computer network , cellular network , quality of experience , network congestion , visibility , process (computing) , tree (set theory) , network performance , reduction (mathematics) , distributed computing , quality of service , real time computing , network packet , mathematical analysis , physics , mathematics , optics , operating system , geometry
Over last 4–5 years, there has been an exponential increase in the amount of traffic transmitted over cellular network. Importantly, with humongous increase in the Wi‐Fi hotspots across many places, subscribers tend to manually switch from cellular to Wi‐Fi network. However, in this process, the operator loses the visibility of subscribers, and is not able to control the congestion in network, and importantly, is not able to provide a guaranteed quality of experience (QoE) to the subscribers. In this study, a network‐centric near‐real‐time mechanism is proposed for maintaining QoE among the subscribers but is found to be a non‐deterministic problem. Importantly, an adaptive dichotomic classification based mechanism is proposed and investigated for offloading the cellular subscribers to Wi‐Fi network, while keeping the computational complexity to less than O (log( n )), where n is the number of nodes in a binary tree. Further, an algorithm is proposed for intelligent selection of subscriber profile that would enable faster completion of the offloading process. An extensive simulation of the proposed algorithm and comparison with state‐of‐the‐art mechanisms reveal up to seven time reduction in computational time; with additional performance increase occurring with increasing number of subscribers.