The uncontrolled growth of Internet of Things (IoT) gadgets have rendered the traditional authentication protocols inadequate, due to the imminent limitation of computer power, latency, and energy usage. The study is a rigorous research on lightweight and blockchain-based authentication schemes by synthesizing the results of thirty-five peer-reviewed articles to evaluate the cryptographic effectiveness in a restricted environment. Strong focus is made on the methods of integrating elliptic curve cryptography, zero-knowledge proofs, Merkle hash trees, and consensus algorithms in distributed ledgers. Systematic benchmarking is done in performance metrics like authentication latency, computational overhead, energy consumption, and rate of mitigating attacks. Lightweight schemes have a latency improvement of up to 68 percent and an energy consumption greater than 54 percent, but blockchain-based schemes achieve a 76 percent improvement on trust verifiability and an average resistance to replay attacks of 92 percent. Mutual authentication robustness and scalability in vehicular networks, smart grids and fog based systems are analyzed using a comparative framework. The combination of light-weight computation and decentralization of trust results in a hybridised model that has an ideal throughput and does not compromise confidentiality or anonymity. It is notable that hybrid protocols show less than 2.1% mean packet loss rate in dynamic topology. The paper provides a comprehensive performance based classification of IoT authentication techniques that ends up with an elegant roadmap towards future generation secure communication systems.