Processing nearest neighbour queries is a fundamental issue that appears in a variety of industries, including machine learning and geographic databases. The Secure Nearest Neighbor (SNN) problem in cloud computing is the focus of this study. Prior SNN schemes are ineffective and insecure. In this study, we officially establish and empirically show that the SNN scheme ASPE is genuinely vulnerable to attacks that use solely ciphertext. Although earlier research demonstrated that it is difficult to build an SNN method even in significantly permissive standard security models, we highlight the shortcomings of the hardness proof. We provide an SNN architecture and demonstrate its resistance to adaptive chosen keyword attacks. Our system is effective because the complexity of processing queries is exponential. To analyse the efficiency of our SNN scheme, we developed our scheme in C++ and compared its performance with a plain text scheme, binary scheme, and a PIR scheme on a huge collection of over 10 million real-world data points. Experimental results show that our scheme is fast (0.124 millisecond per query when data set size is 10 million) and scalable in terms of the number of data points.