Design of a lightweight cryptography-based secure communication model for the internet of things

The Internet of Things describes physical objects with technologies that have sensors, processing capability, and exchange data with other devices (Through internet or other communication networks). Wireless sensor networks, as a subset of IoT networks, consist of nodes that sense, collect and transmit data. The high volume of data detected and transmitted among nodes in these networks makes it necessary to keep data secure and untouched. Since in an IoT network, nodes have restricted resources, implementing typical encryption algorithms are costly and inefficient. Many solutions meet the security demands of resource-constrained wireless sensor networks, and lightweight block ciphers are sophisticated solutions for providing end-to-end security. The aim of this thesis study is to present a secure communication mechanism developed for wireless sensor networks, taking the application type, security level and bit error rate into account. To be used in light weight encryption studies; Performance values of AES, PRESENT, LBlock, Skipjack, SIMON, XTEA, PRINCE, Piccolo, HIGHT and RECTANGLE lightweight algorithms were tested and compared for both encryption and decryption modes in terms of key factors such as memory usage (RAM and ROM), energy and power consumption, throughput and execution time. In the selection of the mentioned algorithms, the criteria of block length, key sizes, popularity, internal structures and application possibilities on software environments were taken into consideration. Some algorithms, such as PRESENT, were hardware-oriented in the early versions, but were later developed to be installed on software platforms. Except for AES which has a block length of 128 bits, all other algorithms have 64 bits block lengths and a key size of 80 or 128 bits. Data transmission among devices has been done over the Dropbox cloud for encryption and decryption modes. Performance analysis was done using Raspberry Pi 3 and Arduino Mega 2560 in IoT environment. An improved security mechanism is also presented for wireless sensor networks, taking the application, security level and bit error rate into account. Reserved frame control field bits are used in the Zigbee MAC header to give the user the ability to choose between insecure or secure communication modes. Under certain conditions, using the cross-layer interaction mechanism the network can switch from AES to other available algorithms (RECTANGLE, Fantomas, and Camellia) due to special criteria. Camellia is certified by the ISO/IEC organization and has comparable security levels to AES. RECTANGLE and Fantomas have high-speed software applications using the bit-slice technique, which makes them a good choice for online applications. AES performance was measured and compared with the performance of the selected algorithms based on memory usage, efficiency, battery consumption, and end-to-end latency for different scenarios in hardware (MICAz mote) and simulator (Riverbed Modeler) in Windows 10. In addition, to make an end-to-end encryption over MQTT-SN messaging protocol, using the 3 reserved MsgType bit fields in the MQTT-SN message packet format, the three algorithms (RECTANGLE, Fantomas, and Camellia) were defined in the MQTT-SN message packet and their performances were compared against AES. Performance evaluations were conducted for different scenarios using the Zolertia Z1 mote and the Contiki (Cooja) simulator on Linux. The evaluations were based on metrics such as memory usage, throughput, battery consumption, and end-to-end latency.