Journal of Innovation in Computer Science and Engineering

Distributed Denial of Service (DDoS) attack is a prominent issue in the Internet world. In DDoS attack, attacker attempts to make certain service unavailable to the intended users, by unnecessarily engaging victim resources. In client-Server applications like web sites, server needs to spend some of its resource in order to handle client request, such as access to database or any heavy computational task. Then with few such requests sever may gets stuck and leads to denial of service attack. We proposed a framework to handle such kind of DDoS attacks. In our pro-posed strategy, when server reaches certain threshold limit (precomputed) of their resources then our solution get invokes and sends a challenge of some NP problem to solve, which can be generate in a negligible amount of time, to the requesting clients. Then the server will respond or process that request only if the client compute and send the solution to that instance of NP problem challenge. Otherwise server drops the connection and save its resources. As compare to previous work, the server side load is reduced and thereby it can handle more number of clients with this approach.