17 January 2020 in IEEE Internet of Things Journal
Internet of Things (IoT) refers to the concept of enabling Internet connectivity and associated services to non-traditional computers formed by integrating essential computing and communication capability to physical things for everyday usage. Security and privacy are two of the major challenges in IoT. The essential security requirements of IoT cannot be ensured by the existing security frameworks due to the constraints in CPU, memory, and energy resources of the IoT devices. Also, the centralized security architectures are not suitable for IoT because they are subjected to single point of attacks. Defending against targeted attacks on centralized resources is expensive. Therefore, the security architecture for IoT needs to be decentralized and designed to meet the limitations in resources. Blockchain is a decentralized security framework suitable for a variety of applications. However, blockchain in its original form is not suitable for IoT, due to its high computational complexity and low scalability. In this paper, we propose a sliding window blockchain (SWBC) architecture that modifies the traditional blockchain architecture to suit IoT applications. The proposed sliding window blockchain uses previous (n-1) blocks to form the next block hash with limited difficulty in Proof-of-Work. The performance of SWBC is analyzed on a real-time data stream generated from a smart home testbed. The results show that the proposed blockchain architecture increases security and minimizes memory overhead while consuming fewer resources.
19 March 2025 in Computer Networks
14 September 2020 in IEEE Transactions on Network and Service Management
A novel framework, Medium-Term Disruption Tolerant Software Defined Network (MDT-SDN), is proposed to handle medium-term disruptions of the order 10 seconds to 6 minutes. Such medium-term disruptions are crucial for the next generation wireless networks that use TCP/IP protocol stack, where the existing disruption tolerant approaches may under-perform due to resource constraints. MDT-SDN enables network control with an additional STORE action which exploits the nodes’ memory for buffering packets within the TCP/IP stack during link disruptions and forwards them as the links become alive. First, we model the network with medium-term disruptions using temporal graphs and design an Earliest Arrival Path with Minimal Storage Time (EAPMST) controller algorithm to demonstrate the MDT-SDN framework. The framework is realized in a wireless mesh network testbed, using in-band control, involving mobile nodes with three mobility models: ( i) stationary nodes, ( ii) group mobility, and ( iii) random mobility. Experimental results with four application protocols justify the efficacy of MDT-SDN in providing storage control and additional performance benefits. MDT-SDN along with EAPMST improves the throughput beyond 25% with random mobility model and is capable of carrying packets and maintaining sessions during medium-term disruptions using the existing TCP/IP stack. The use of EAPMST reduces packets’ buffering time by 24.45% in group mobility while the buffering time increases in random mobility by 5.82%. However, EAPMST significantly reduces the buffer occupancy by 60.28% and 44.49% for group and random mobility, respectively.