Akademik Veri Yönetim Sistemi
Tezin Türü: Yüksek Lisans
Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü, Türkiye
Tezin Onay Tarihi: 2018
Öğrenci: DOĞANALP ERGENÇ
Danışman: ERTAN ONUR
Özet:Separation of user (data) plane from control plane in networks helps scale resources independently, increase the quality of service and facilitate autonomy. In ad-hoc networks, the plane-separation through clustering introduces a hierarchy where control functions can be carried out by some designated cluster heads and other nodes perform according to the outcomes of those functions. Therefore, clustered topologies can be considered as a natural consequence of the control and user plane separation (CUPS). Moreover, hierarchical routing protocols, which are constructed upon the clustered topologies, enable the use of CUPS architecture for the end-to-end communication. However, there is no silver bullet to apply clustering algorithms that are directly dependent on the network characteristics, and the routing protocols designed for clustered topologies cannot effectively utilize CUPS since they neglect the role of the nodes in the data plane. This study investigates the application of CUPS architecture in ad-hoc networks by considering clustering and routing protocols holistically. First, the adaptability of the clustering techniques is discussed to satisfy different objectives such as stability, energy efficiency and service quality; and Dependability-based Clustering Algorithm (DCA) is proposed. DCA is a dynamic clustering algorithm that exploits a cross-layer architecture. Its different parameters are analyzed and optimized using the sensitivity analysis technique, Moment-Independent Delta Analysis. Then, the hierarchical routing protocol CUPS-based Hierarchical Routing Algorithm (CHRA) is proposed for end-to-end communication. In CHRA, the separate functions of the control and data planes are explicitly defined to provide the quality of service and energy efficiency taking advantage of the clustered topology. The overall CUPS-centric framework including DCA and CHRA is implemented in the discrete event-based simulator, OMNeT++. The results show that DCA outperforms its opponents when it is optimized for different scenarios. Besides, the study reveals the significant points that are required to be considered for designing clustering algorithms through the discussion of the optimization process. Finally, CHRA offers a better quality of service and a fair energy consumption thanks to its novel approach that considers the effective use of the data plane as well as the control plane. The complete plane-separated approach is utilized for energy efficiency and the quality of service in ad-hoc networks.