Study of named data networking algorithms in internet of things

Study of Named Data Networking Algorithms in Internet of Things. The main protocol of the Internet – TCP/IP is based on an end-to-end delivery design. It remains in the Internet protocol stack, though a vast majority of users are much more interested in a data itself than a reliable data delivery. It implies that an effective search and publishing of content should be a fundamental goal of the global network. It is expected a billion of small battery-powered devices, which have a lack of a memory resource and a processor capacity, to join the Internet. The Internet of these constrained devices – the Internet of Things will consist of unreliable and lossy channels due to resource limitations and constant interference. Information-centric networking (ICN) technologies use content names, which describe the data that must be sent. In this way, an application layer becomes much more involved in data delivery. Named data networking (NDN) technology goes along with a main principle of ICN – the usage of content names. Besides, it stands out for its unique forwarding strategy layer and a decentralized distributed caching. Using traditional NDN architecture every group data element can be retrieved only by sending a separate interest packet. This way vulnerable network and energy resources are wasted and critical information is detained. The main purpose of this work is to create NDN algorithms that enable multi-source data retrieval on a single request. The primary tasks are to rethink and implement interest and data pipelines and design new data-cache search algorithm, which would be able to find all data elements matching the group request. After comprehensive NDN architecture analysis interest and data pipelines were reformed and rethought data-cache search algorithm was created by modifying NFD source code. ndnSIM simulator was used to create a network scenario. Simulation results revealed that newly implemented NDN algorithms allow multi-source data retrieval using a single group data request and ensure that all matching data elements are found in the cache. It was observed that our methods can significantly reduce the number of interest packets sent by client, but the network is overwhelmed with redundant data packets and algorithm optimization is needed. Simulations results also showed that, in case of exact prefix match, interest and data packets are managed in traditional NDN manner