IoT is a network of multiple electronic smart devices which are connected over the internet sharing every kind of detail including sending and receiving of the data over a common interconnected network and it includes computation systems, electrical systems, etc. IoT supports a lot of applications which include medical healthcare, smart cities, smart devices, smart homes, smart vehicles, public safety, etc. With the increase in the technological advancements, there is an increase in the security issues in the IoT network. Various kinds of threats emerge everyday in different applications, architectures, etc. Network security is major issue in this and such threats include malicious activities, DDoS attacks, spying of data, etc. are the merging issues in the IoT devices. This paper proposes a framework which can be useful in the coming future as it uses optimized multi kernel based model using machine learning technique. The performance is measured with the regular RSA algorithm and already existing algorithms over several parameters like computation time, compile time, etc.