Security and privacy aspects of mobile platforms and applications security

Mobile smart devices (such as smartphones and tablets) emerged to dominant computing platforms for end-users. The capabilities of these convenient mini-computers seem nearly boundless: They feature compelling computing power and storage resources, new interfaces such as Near Field Communication (NFC) and Bluetooth Low Energy (BLE), connectivity to cloud services, as well as a vast number and variety of apps. By installing these apps, users can turn a mobile device into a music player, a gaming console, a navigation system, a business assistant, and more. In addition, the current trend of increased screen sizes make these devices reasonable replacements for traditional (mobile) computing platforms such as laptops. On the other hand, mobile platforms process and store the extensive amount of sensitive information about their users, ranging from the user’s location data to credentials for online banking and enterprise Virtual Private Networks (VPNs). This raises many security and privacy concerns and makes mobile platforms attractive targets for attackers. The rapid increase in number, variety and sophistication of attacks demonstrate that the protection mechanisms offered by mobile systems today are insufficient and improvements are necessary in order to make mobile devices capable of withstanding modern security and privacy threats. This dissertation focuses on various aspects of security and privacy of mobile platforms. In particular, it consists of three parts: (i) advanced attacks on mobile platforms and countermeasures; (ii) online authentication security for mobile systems, and (iii) secure mobile applications and services. Specifically, the first part of the dissertation concentrates on advanced attacks on mobile platforms, such as code re-use attacks that hijack execution flow of benign apps without injecting malicious code, and application-level privilege escalation attacks that allow malicious or compromised apps to gain more privileges than were initially granted. In this context, we develop new advanced code re-use attack techniques that can bypass deployed protection mechanisms (e.g., Address Space Layout Randomization (ASLR)) and cannot be detected by any of the existing security tools (e.g., return address checkers). Further, we investigate the problem of application-level privilege escalation attacks on mobile platforms like Android, study and classify them, develop proof of concept exploits and propose countermeasures against these attacks. Our countermeasures can mitigate all types of application-level privilege escalation attacks, in contrast to alternative solutions proposed in literature. In the second part of the dissertation we investigate online authentication schemes frequently utilized by mobile users, such as the most common web authentication based upon the user’s passwords and the recently widespread mobile 2-factor authentication (2FA) which extends the password-based approach with a secondary authenticator sent to a user’s mobile device or generated on it (e.g, a One-time Password (OTP) or Transaction Authentication Number (TAN)). In this context we demonstrate various weaknesses of mobile 2FA schemes deployed for login verification by global Internet service providers (such as Google, Dropbox, Twitter, and Facebook) and by a popular Google Authenticator app. These weaknesses allow an attacker to impersonate legitimate users even if their mobile device with the secondary authenticator is not compromised. We then go one step further and develop a general attack method for bypassing mobile 2FA schemes. Our method relies on a cross-platform infection (mobile-to-PC or PC-to-mobile) as a first step in order to compromise the Personal Computer (PC) and a mobile device of the same user. We develop proof-of-concept prototypes for a cross-platform infection and show how an attacker can bypass various instantiations of mobile 2FA schemes once both devices, PC and the mobile platform, are infected. We then deliver proof-of-concept attack implementations that bypass online banking solutions based on SMS-based TANs and visual cryptograms, as well as login verification schemes deployed by various Internet service providers. Finally, we propose a wallet-based secure solution for password-based authentication which requires no secondary authenticator, and yet provides better security guaranties than, e.g., mobile 2FA schemes. The third part of the dissertation concerns design and development of security sensitive mobile applications and services. In particular, our first application allows mobile users to replace usual keys (for doors, cars, garages, etc.) with their mobile devices. It uses electronic access tokens which are generated by the central key server and then downloaded into mobile devices for user authentication. Our solution protects access tokens in transit (e.g., while they are downloaded on the mobile device) and when they are stored and processed on the mobile platform. The unique feature of our solution is offline delegation: Users can delegate (a portion of) their access rights to other users without accessing the key server. Further, our solution is efficient even when used with constraint communication interfaces like NFC. The second application we developed is devoted to resource sharing among mobile users in ad-hoc mobile networks. It enables users to, e.g., exchange files and text messages, or share their tethering connection. Our solution addresses security threats specific to resource sharing and features the required security mechanisms (e.g., access control of resources, pseudonymity for users, and accountability for resource use). One of the key features of our solution is a privacy-preserving access control of resources based on FoF Finder (FoFF) service, which provides a user-friendly means to configure access control based upon information from social networks (e.g., friendship information) while preserving user privacy (e.g., not revealing their social network identifiers). The results presented in this dissertation were included in several peer-reviewed publications and extended technical reports. Some of these publications had significant impact on follow up research. For example, our publications on new forms of code re-use attacks motivated researchers to develop more advanced forms of ASLR and to re-consider the idea of using Control-Flow Integrity (CFI). Further, our work on application-level privilege escalation attacks was followed by many other publications addressing this problem. Moreover, our access control solution using mobile devices as access tokens demonstrated significant practical impact: in 2013 it was chosen as a highlight of CeBIT – the world’s largest international computer expo, and was then deployed by a large enterprise to be used by tens of thousands of company employees and millions of customers