Securing Password Authentication for Web-based Applications

TL;DR

This paper identifies vulnerabilities in web password input fields that enable phishing attacks and proposes a new secure protocol with a dedicated credential field to prevent such attacks, supported by analysis and implementation testing.

Contribution

It introduces a novel secure protocol and credential field design that address specific vulnerabilities in web password authentication to prevent phishing.

Findings

Identified a design vulnerability in HTML password fields.

Proposed a secure protocol with a new credential field.

Analyzed deployment issues and tested implementation overheads.

Abstract

The use of passwords and the need to protect passwords are not going away. The majority of websites that require authentication continue to support password authentication. Even high-security applications such as Internet Banking portals, which deploy 2-factor authentication, rely on password authentication as one of the authentication factors. However phishing attacks continue to plague password-based authentication despite aggressive efforts in detection and takedown as well as comprehensive user awareness and training programs. There is currently no foolproof mechanism even for security-conscious websites to prevent users from being directed to fraudulent websites and having their passwords phished. In this paper, we apply a threat analysis on the web password login process, and uncover a design vulnerability in the HTML<inputtype="password"> field. This vulnerability can be exploited for phishing attacks as the web authentication process is not end-to-end secured from each input password field to the web server. We identify four properties that encapsulate the requirements to stop web-based password phishing, and propose a secure protocol to be used with a new credential field that complies with the four properties. We further analyze the proposed protocol through an abuse-case evaluation, discuss various deployment issues, and also perform a test implementation to understand its data and execution overheads