ROUTING PROTOCOL AND ARCHITECTURE FOR SKALABILITY AND MOBILITY SUPPORT IN NAMED DATA NETWORK

• Main Author: Ariefianto Wibowo, Tody
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/76697
• Institution: Institut Teknologi Bandung

Development of internet technology today, has been woried will bring the network to collapse, due to shifting nature of the internet itself since created from nowadays characteristic. Named Data Network (NDN) brings a new paradigm and provides fundamental changes in the implementation of data communications on the internet. Routing is one of the important factors in messaging delivery mechanism on NDN and the scalability in perutean is an issue that arises with the increasing number of entities on the internet. NDN research starts from Simulation which is a method that demands the smallest resources, then continues to expand towards emulation and testbeds. Testbeds with prototype devices are closest to real world reality but are increasingly difficult to implement. A framework is needed that can bring NDN research closer to the real world with more controlled efforts and resources. In the first part, the proposed routing protocol is named Grid-based Scalable NDN Routing (GSNR). Where previously a mathematical model for the efficiency of geographic routing protocols and link state was built on the NDN. In the second part of the research, a framework is proposed that combines the advantages of emulation and testbed. The framework involves the NDN-PnetLab (PL) emulation process using Qemu-based VMs for an easier implementation approach, comprehensive monitoring, and cost-effectiveness. GSNR, by using a hybrid routing protocol, provides better and more stable performance than the previous routing protocols NLSR and GEO. GSNR is able to provide low message overhead without sacrificing data transmission performance. GSNR shows stable performance for throughput and delay, with experiments scenarios, changing the area and the number of nodes. GSNR performance (overhead, path-stretch, delay, throughput), on average, only changed (increased or decreased) 19.83% compared to 116.46% in NLSR and 50.91% in GEO. The results of the observations show how the proposed NDN-PnetLab (PL) framework can monitor QoS parameters, CPU usage, and RAM usage. CPU and RAM usage is impossible to check and analyze if we only use NDNSim or even Mini-NDN. Compared to Mini-NDN, the PL framework provides a performance iv difference of 7.59% on RTT delay, and 0.89% on nCHR, then can provide intensive parameter measurement results related to hardware performance. This is urgently needed to continue research into the implementation of NDN systems.