Main Article Content
Denial of service (DOS) and distributed denial of service (DDOS) have become major threats in a computer network. The function of a DDOS attack is fundamental to make the system not to respond any requests or providing services. Detection and prevention of DOS/DDOS is a tedious process. In the earlier days, it can be prevented by using the puzzlesolving technique. In puzzle-solving technique first, the server generates some puzzles and it sends to the client. Then the client solves the puzzle and it sends it to the server then the server confirms that the client able to obtain the service from the server. In this type of puzzle-solving technique attackers can merely reply to the server it cannot be identified in this puzzle-solving technique and Graphics Processing Unit (GPU)-inflation DoS doesn’t work and it publish puzzle function in advance. It can be enhanced or overcome by using software puzzle-solving technique it dynamically generates puzzle function. This technique exploits the architectural difference between CPU and GPU. If an attacker tries to move a puzzle transition task into CPU either they want to translate into functionally equivalent GPU or it should do at dynamic puzzle generation. It is the timeconsuming process to translate or rewriting a software puzzle. This software puzzle has some drawbacks it cannot carry out puzzle solving process on cloud environment and it doesn’t construct client side software puzzle. In order to overcome this problem here, we proposed a new client puzzle technique which is based on two puzzles between clients. Moreover, in order to improve the performance of Denial of Service attack brute force attack, command evaluation, and Kullback libeller divergence are utilized.