Transition in Monitoring and Network Offloading - Handling Dynamic Mobile Applications and Environments

• Main Author: Richerzhagen, Nils
• Format: Others
• Language: en
• Published: 2019
• Online Access: http://tuprints.ulb.tu-darmstadt.de/8612/13/2019-04-08_Richerzhagen_Nils.pdf; Richerzhagen, Nils <http://tuprints.ulb.tu-darmstadt.de/view/person/Richerzhagen=3ANils=3A=3A.html> : Transition in Monitoring and Network Offloading - Handling Dynamic Mobile Applications and Environments. Technische Universität, Darmstadt [Ph.D. Thesis], (2019)

Communication demands increased significantly in recent years, as evidenced in studies by Cisco and Ericsson. Users demand connectivity anytime and anywhere, while new application domains such as the Internet of Things and vehicular networking, amplify heterogeneity and dynamics of the resource-constrained environment of mobile networks. These developments pose major challenges to an efficient utilization of existing communication infrastructure. To reduce the burden on the communication infrastructure, mechanisms for network offloading can be utilized. However, to deal with the dynamics of new application scenarios, these mechanisms need to be highly adaptive. Gathering information about the current status of the network is a fundamental requirement for meaningful adaptation. This requires network monitoring mechanisms that are able to operate under the same highly dynamic environmental conditions and changing requirements. In this thesis, we design and realize a concept for transitions within network offloading to handle the former challenges, which constitutes our first contribution. We enable adaptive offloading by introducing a methodology for the identification and encapsulation of gateway selection and clustering mechanisms in the transition-enabled service AssignMe.KOM. To handle the dynamics of environmental conditions, we allow for centralized and decentralized offloading. We generalize and show the significant impact of our concept of transitions within offloading in various, heterogeneous applications domains such as vehicular networking or publish/subscribe. We extend the methodology of identification and encapsulation to the domain of network monitoring in our second contribution. Our concept of a transition-enabled monitoring service AdaptMon.KOM enables adaptive network state observation by executing transitions between monitoring mechanisms. We introduce extensive transition coordination concepts for reconfiguration in both of our contributions. To prevent data loss during complex transition plans that cover multiple coexisting transition-enabled mechanisms, we develop the methodology of inter-proxy state transfer. We target the coexistence of our contributions for the use case of collaborative location retrieval on the example of location-based services. Based on our prototypes of AssignMe.KOM and AdaptMon.KOM, we conduct an extensive evaluation of our contributions in the Simonstrator.KOM platform. We show that our proposed inter-proxy state transfer prevents information loss, enabling seamless execution of complex transition plans that cover multiple coexisting transition-enabled mechanisms. Additionally, we demonstrate the influence of transition coordination and spreading on the success of the network adaptation. We manifest a cost-efficient and reliable methodology for location retrieval by combining our transition-enabled contributions. We show that our contributions allow for adaption on dynamic environmental conditions and requirements in network offloading and monitoring.