Abstract:
IoT has the potential to affect our ways of life. It is the next step of Internet where
all the physical objects around us will be connected to each other. According to
Gartner, by 2020 there will be over 26 billion connected devices. However, the
security of such a big network of interconnected devices is of paramount
importance. According to a report from Russian-based Kaspersky Labs, botnets -
not spam, viruses or worm- currently pose the biggest threat to the
Internet.However, few works that have been done on this issue in the recentpast are
not successful on themselves alone. In this paper, I present an in depth
understanding of the problem and propose a mechanism to counter this issue. My
proposal is based on Transductive Confidence Machines, which was previously
proposed as a mechanism to provide confidence measures on classification
decisions. It proposes to make use of this algorithm with the help of honeypot to
collect attack data and uses these data to make the system more proficient.
