The integration of computing, communication and sensor having self-configuring features is what we call IoT. Its span a variety of application area like smart cities, healthcare, industrial automation, transportation, etc. However, these complex application areas make the system more vulnerable to threats like Dos attack, sybil attack, various types of network attacks leading to data breaches and unauthorized access that can compromise the integrity of information and privacy of users. In order to address these threats, integration of technologies like blockchain and machine learning with IoT can solve the problem up to a great extent, particularly machine learning helps to identify and detects the threats before attacks, and blockchain improves the communication safety enabling devices to communicate in a complex system. This paper outlines IoT systems, its pillars and architectural layer, this paper also highlighted various types of threats in IoT systems like application attacks, DoS attack, sybil attacks etc., further the paper provides various effective measures to tackle these attacks by incorporating various domain like machine learning, Blockchain and Software-Defined Networking. Finally, the paper is concluded highlighting several recent advancements in this direction that can further enhance the security framework.