Detection of Lightweight Directory Access Protocol Query Injection Attacks in Web Applications

The Lightweight Directory Access Protocol (LDAP) is a common protocol used in organizations for Directory Service. LDAP is popular because of its features such as representation of data objects in hierarchical form, being open source and relying on TCP/IP, which is necessary for Internet access. However, with LDAP being used in a large number of web applications, different types of LDAP injection attacks are becoming common. The idea behind LDAP injection attacks is to take advantage of an application not validating inputs before being used as part of LDAP queries. An attacker can provide inputs that may result in alteration of intended LDAP query structure. LDAP injection attacks can lead to various types of security breaches including (i) Login Bypass, (ii) Information Disclosure, (iii) Privilege Escalation, and (iv) Information Alteration. Despite many research efforts focused on traditional SQL Injection attacks, most of the proposed techniques cannot be suitably applied for mitigating LDAP injection attacks due to syntactic and semantic differences between LDAP and SQL queries. Many implemented web applications remain vulnerable to LDAP injection attacks. In particular, there has been little attention for testing web applications to detect the presence of LDAP query injection attacks. The aim of this thesis is two folds: First, study various types of LDAP injection attacks and vulnerabilities reported in the literature. The planned research is to critically examine and evaluate existing injection mitigation techniques using a set of open source applications reported to be vulnerable to LDAP query injection attacks. Second, propose an approach to detect LDAP injection attacks by generating test cases when developing secure web applications. In particular, the thesis focuses on specifying signatures for detecting LDAP injection attack types using Object Constraint Language (OCL) and evaluates the proposed approach using PHP web applications. We also measure the effectiveness of generated test cases using a metric named Mutation Score.