A new features vector matching for big heterogeneous data in intrusion detection context

Nowadays, the volume of data considerably increasing, the data is exploding on the scale of the Exabyte and the Zettabyte at an exceptionally high rate. These can be characterized as big data. Hence, the security of the network, Internet, websites, Iot devices and the organizations, of this growth is indispensable. Detecting intrusions in such a big heterogeneous data environment is challenging. In this paper, we will present a new representation of data that can support this big heterogeneous environment. We will use three different datasets and propose an automatically matching algorithm that measures the semantic similarity between each two features existing on different datasets. Thereafter, an approximate vector is created that any type of coming data can be stored. With this representation, we can have subsequently an efficient intrusion detection system that can be able to acknowledge any instance of the existing data in the networks.

I. Introduction

Technological evolution in networking, Internet, the Iot devices, the 5G communication media and the worldwide network traffic lead a huge amount of data to be generated every second from heterogeneous sources.