No abstract

No abstract

The Internet of Things (IoT) extends the Internet wherein real world things are part of a computing network. The IoT has seen exponential growth and according to Cisco prediction about 50 billion devices will be connected by 2020. Handling this massive scale is challenging research issue. In this paper, we use cloud-fog-edge based IoT middleware for distributed IoT service provisioning. We model IoT middleware using queueing network, perform analytical analysis of IoT middleware components. Followed by dynamic scaling algorithm which considers contention and coherency as limiting factors for scalability. It is used for quantitative analysis of IoT middleware with the increasing workload. The scalability function was evaluated using the simulation for important performance and scalability parameters namely throughput, CPU utilization and response time. It is observed that because of contention and coherency overhead, the proposed approach is able to scale sub-linearly which is practical compared ideal scalability of multi-server queueing network and not very restrictive as given universal scalability law(USL) applied for tightly coupled systems.

Neural cryptography is based on synchronization of tree parity machines by mutual learning. We extend previous key-exchange protocols by replacing random inputs with queries depending on the current state of the neural networks. The results show that queries restore the security against attackers. We further restrict amount information available to an attacker by keeping seed of pseudo random number generator private.

No abstract