Blok zinciri ile nesnelerin interneti güvenliği

The Internet of things (IoT) is one of the pioneer concepts that is shaping the "smart" world of today and the future. It is a core enabler of advancements in many fields, from agriculture to astronomy, industrial automation to autonomous vehicles, etc. However, most IoT deployments include cost-efficient lightweight devices with limited resources. Hence, the entire network must be built in the simplest way. This necessity forces engineers to include more sophisticated devices in the network, such as gateways and servers. So that, although the nodes are widely distributed in a geographical or topological sense, the network turns into a centralized structure that eventually creates bottlenecks and single-points-of-failure. Among the issues caused by that phenomenon, we focused on solving the data manipulation and event management problems. In a nutshell, in most IoT scenarios, data flows from sensor devices to data storage and processing units. Contrarily, event-driven commands flow from these units to actuator devices, if any. Therefore, an attacker who gained access to the centralized units can leak, alter, or even remove critical data and may exploit event handling features. This is where blockchain, another revolutionary technology, may be extremely handy. Integration of a decentralized ledger as a data storage provides data integrity, immutability, and non-repudiation for an IoT deployment. Moreover, the adoption of a custom smart contract lets the IoT deployment benefit from decentralized and immutable event management features, too. Our thesis introduces a novel IoT architecture scheme that incorporates an Ethereum-based private blockchain that runs an ad-hoc smart contract in an MQTT-based IoT deployment containing sensor and actuator nodes to provide the functionalities. Apart from the architectural design, the thesis also presents a simulation-based proof-of-concept built for validation purposes, as well as performance measurements. Per our benchmarks, the proposed scheme is shown to be valid, scalable, and efficient for various IoT scenarios.