Secure remote user authentication for multi-server environment using machine learning technique

The earlier remote password authentication schemes required a service providing server to authenticate a legitimate user for remote login. However, the traditional schemes are not useful in multi-server architecture because of multiple user ids and passwords. In this paper, we present a remote password authentication scheme for multi-server architecture that can be robust and improved network security. This password authentication system is a trained classification system based on SVM. In this scheme, the users only remember own identity and password and own choices for the various server login which he filled during registration where the user can choose his password at will. Furthermore, this system does not require having any overhead of password or verification table and is very dynamic in nature and can also withstand the replay attack and masquerading.