Nuisance calls are not only irritating and annoying but also pose several security threats. Nuisance calls can be classified into live, recorded, scam and silent spam calls. Unlike traditional communication networks, users identities over the web cannot be asserted as they are created without any subscription contract or identification proof. To address this issue, we present N-Combat framework that effectively combats automated and manually generated nuisance calls over web communication services. In N-Combat the caller's reputation is computed using total call duration, out-degree, feedback received and the reliability of communicating participants. To address the dynamic behavior of callers, the concept of a dual time window is introduced. Furthermore, the framework uses a watchdog mechanism to combat whitewashing attacks that allows spammers to shed their bad reputation and reenter the network with a new identity