Impact of Time in Network Selection for Mobile Nodes

Today, mobile nodes use multiple Internet access networks inefficiently. State-of-the-art network selection strategies distribute data traffic to available networks, but ignore an important second dimension: time. Time selection offers the opportunity to plan usage of future-available networks for delay-tolerant data traffic. We hypothesize, that concurrent selection of network and time leads to synergy effects, which reduce transmission cost and boost connectivity performance. To assess data distribution to wireless networks and time, we propose a novel rating model for joint network and time selection. The proposed model rates the satisfaction of Quality-of-Service (QoS) application requirements and trades off conflicting optimization goals. Moreover, we analyze the impact of time in network selection and present three network selection schedulers, which differ in their time selection strategy. Evaluation of the results reveals a strong impact of time selection on network performance. This gives evidence, that our initial hypothesis holds and forward-looking scheduling strategies provide a substantial benefit over state-of-the-art approaches.