Musard Balliu

Epistemic temporal logic for information flow security

Temporal epistemic logic is a well-established framework for expressing agents knowledge and how it evolves over time. Within language-based security these are central issues, for instance in the context of declassification. We propose to bring these two areas together. The paper presents a computational model and an epistemic temporal logic used to reason about knowledge acquired by observing program outputs. This approach is shown to elegantly capture standard notions of noninterference and declassification in the literature as well as information flow properties where sensitive and public data intermingle in delicate ways.

Automating Information Flow Analysis of Low Level Code

Low level code is challenging: It lacks structure, it uses jumps and symbolic addresses, the control flow is often highly optimized, and registers and memory locations may be reused in ways that make typing extremely challenging. Information flow properties create additional complications: They are hyperproperties relating multiple executions, and the possibility of interrupts and concurrency, and use of devices and features like memory-mapped I/O requires a departure from the usual initial-state final-state account of noninterference. In this work we propose a novel approach to relational verification for machine code. Verification goals are expressed as equivalence of traces decorated with observation points. Relational verification conditions are propagated between observation points using symbolic execution, and discharged using first-order reasoning. We have implemented an automated tool that integrates with SMT solvers to automate the verification task. The tool transforms ARMv7 binaries into an intermediate, architecture-independent format using the BAP toolset by means of a verified translator. We demonstrate the capabilities of the tool on a separation kernel system call handler, which mixes hand-written assembly with gcc-optimized output, a UART device driver and a crypto service modular exponentiation routine.

ENCoVer: Symbolic Exploration for Information Flow Security

We address the problem of program verification for information flow policies by means of symbolic execution and model checking. Noninterference-like security policies are formalized using epistemic logic. We show how the policies can be accurately verified using a combination of concolic testing and SMT solving. As we demonstrate, many scenarios considered tricky in the literature can be solved precisely using the proposed approach. This is confirmed by experiments performed with ENCOVER, a tool based on Java Path Finder and Z3, which we have developed for epistemic noninterference concolic verification.

Explicit Secrecy: A Policy for Taint Tracking

Taint tracking is a popular security mechanism for tracking data-flow dependencies, both in high-level languages and at the machine code level. But despite the many taint trackers in practical use, the question of what, exactly, tainting means - what security policy item bodies - remains largely unexplored. We propose explicit secrecy, a generic framework capturing the essence of explicit flows, i.e., the data flows tracked by tainting. The framework is semantic, generalizing previous syntactic approaches to formulating soundness criteria of tainting. We demonstrate the usefulness of the framework by instantiating it with both a simple high-level imperative language and an idealized RISC machine. To further understanding of what is achieved by taint tracking tools, both dynamic and static, we obtain soundness results with respect to explicit secrecy for the tainting engine cores of a collection of popular dynamic and static taint trackers.

A weakest precondition approach to active attacks analysis

Information flow controls can be used to protect both data confidentiality and data integrity. The certification of the security degree of a program that runs in untrusted environments still remains an open problem in language-based security. The notion of robustness asserts that an active attacker, who can modify program code in some fixed points (holes), is not able to disclose more private information than a passive attacker, who merely observes public data. In this paper, we extend a method recently proposed for checking declassified non-interference in presence of passive attackers only, in order to check robustness by means of the weakest precondition semantics. In particular, this semantics simulates the kind of analysis that can be performed by an attacker, i.e., from the public output towards the private input. The choice of the semantics lets us distinguish between different attacks models. In this paper, we also introduce relative robustness that is a relaxed notion of robustness for restricted classes of attacks.

Formalising privacy policies in social networks

Social Network Services (SNS) have changed the way people communicate, bringing many benefits but also new concerns. Privacy is one of them. We present a framework to write privacy policies for SNSs and to reason about such policies in the presence of events making the network evolve. The framework includes a model of SNSs, a logic to specify properties and to reason about the knowledge of the users (agents) of the SNS, and a formal language to write privacy policies. Agents are enhanced with a reasoning engine allowing the inference of knowledge from previously acquired knowledge. To describe the way SNSs may evolve, we provide operational semantics rules which are classified into four categories: epistemic, topological, policy, and hybrid, depending on whether the events under consideration change the knowledge of the SNS users, the structure of the social graph, the privacy policies, or a combination of the above, respectively. We provide specific rules for describing Twitters behaviour, and prove that it is privacy-preserving (i.e., that privacy is preserved under every possible event of the system). We also show how Twitter and Facebook are not privacy-preserving in the presence of additional natural privacy policies.

A weakest precondition approach to robustness

With the increasing complexity of information management computer systems, security becomes a real concern. E-government, web-based financial transactions or military and health care information systems are only a few examples where large amount of information can reside on different hosts distributed worldwide. It is clear that any disclosure or corruption of confidential information in these contexts can result fatal. Information flow controls constitute an appealing and promising technology to protect both data confidentiality and data integrity. The certification of the security degree of a program that runs in untrusted environments still remains an open problem in the area of language-based security. Robustness asserts that an active attacker, who can modify program code in some fixed points (holes), is unable to disclose more private information than a passive attacker, who merely observes unclassified data. In this paper, we extend a method recently proposed for checking declassified non-interference in presence of passive attackers only, in order to check robustness by means of weakest precondition semantics. In particular, this semantics simulates the kind of analysis that can be performed by an attacker, i.e., from public output towards private input. The choice of semantics allows us to distinguish between different attacks models and to characterize the security of applications in different scenarios. n nOur results are sound to address confidentiality and integrity of software running in untrusted environments where different actors can distrust one another. For instance, a web server can be attacked by a third party in order to steal a session cookie or hijack clients to a fake web page.

JSLINQ: Building Secure Applications across Tiers

Modern web and mobile applications are complex entities amalgamating different languages, components, and platforms. The rich features span the application tiers and components, some from third parties, and require substantial efforts to ensure that the insecurity of a single component does not render the entire system insecure. As of today, the majority of the known approaches fall short of ensuring security across tiers. This paper proposes a framework for end-to-end security, by tracking information flow through the client, server, and underlying database. The framework utilizes homogeneous meta-programming to provide a uniform language for programming different components. We leverage .NET meta-programming capabilities from the F# language, thus enabling language-integrated queries on databases and interoperable heterogeneous execution on the client and the server. We develop a core of our security enforcement in the form of a security type system for a functional language with mutable store and prove it sound. Based on the core, we develop JSLINQ, an extension of the WebSharper library to track information flow. We demonstrate the capabilities of JSLINQ on the case studies of a password meter, two location-based services, a movie rental database, an online Battleship game, and a friend finder app. Our experiments indicate that JSLINQ is practical for implementing high-assurance web and mobile applications.

A Logic for Information Flow Analysis of Distributed Programs

Securing communication in large scale distributed systems is an open problem. When multiple principals exchange sensitive information over a network, security and privacy issues arise immediately. ...

We are family: Relating information-flow trackers

While information-flow security is a well-established area, there is an unsettling gap between heavyweight information-flow control, with formal guarantees yet limited practical impact, and lightweight tainting techniques, useful for bug finding yet lacking formal assurance. This paper proposes a framework for exploring the middle ground in the range of enforcement from tainting (tracking data flows only) to fully-fledged information-flow control (tracking both data and control flows). We formally illustrate the trade-offs between the soundness and permissiveness that the framework allows to achieve. The framework is deployed in a staged fashion, statically embedding a dynamic monitor, being parametric in security policies, as they do not need to be fixed until the final deployment. This flexibility facilitates a secure app store architecture, where the static stage of verification is performed by the app store and the dynamic stage is deployed on the client. To illustrate the practicality of the framework, we implement our approach for a core of Java and evaluate it on a use case with enforcing privacy policies in the Android setting. We also show how a state-of-the-art dynamic monitor for JavaScript can be easily adapted to implement our approach.