Memory safety is traditionally characterized in terms of bad things that cannot happen. This approach is currently embraced in the literature on formal methods for memory safety. However, a general semantic principle for memory safety, that implies the negative items, remains elusive. This paper focuses on the allocator-specific aspects of memory safety, such as null-pointer dereference, use after free, double free, and heap overflow. To that extent, we propose a notion of gradual allocator independence that accurately captures the allocator-dependent aspects of memory safety. Our approach is inspired by the previously suggested connection between memory safety and noninterference, but extends that connection in a fundamentally important direction towards downgrading. We consider a low-level language with access to an allocator that provides malloc and free primitives in a flat memory model. Pointers are just integers, and as such it is trivial to write memory-unsafe programs. The basic intuition of gradual allocator independence is that of noninterference, namely that allocators must not influence program execution. This intuition is refined in two important ways that account for the allocators running out-of-memory and for programs to have pointer-to-integer casts. The key insight of the definition is to treat these extensions as forms of downgrading and give them satisfactory technical treatment using the state-of-the-art information flow machinery.

Mainstream messaging platforms offer a variety of features designed to enhance user privacy, such as password-protected chats and end-to-end encryption (E2EE), which primarily protect message contents. Beyond contents, a lot can be inferred about people simply by tracing who sends and receives messages, when, and how often. This paper explores user perceptions of and attitudes toward 'untraceability', defined as preventing third parties from tracing who communicates with whom, to inform the design of privacy-enhancing technologies and untraceable communication protocols. Through a vignette-based qualitative study with 189 participants, we identify a diverse set of features that users perceive to be useful for untraceable messaging, ranging from using aliases instead of real names to VPNs. Through a reflexive thematic analysis, we uncover three overarching attitudes that influence the support or rejection of untraceability in messaging platforms and that can serve as a set of new privacy personas: privacy fundamentalists, who advocate for privacy as a universal right; safety fundamentalists, who support surveillance for the sake of accountability; and optimists, who advocate for privacy in principle but also endorse exceptions in idealistic ways, such as encryption backdoors. We highlight a critical gap between the threat models assumed by users and those addressed by untraceable communication protocols. In particular, many participants understood untraceability as a form of anonymity, but interpret it as senders and receivers hiding their identities from each other, rather than from external network observers. We discuss implications for the design of strategic communication and user interfaces of untraceable messaging protocols, and propose framing untraceability as a form of 'altruistic privacy', i.e., adopting privacy-enhancing technologies to protect others, as a promising strategy to foster broader adoption.

Transport layer data leaks metadata unintentionally - such as who communicates with whom. While tools for strong transport layer privacy exist, they have adoption obstacles, including performance overheads incompatible with mobile devices. We posit that by changing the objective of metadata privacy for all traffic, we can open up a new design space for pragmatic approaches to transport layer privacy. As a first step in this direction, we propose using techniques from information flow control and present a principled approach to constructing formal models of systems with metadata privacy for some, deniable, traffic. We prove that deniable traffic achieves metadata privacy against strong adversaries- this constitutes the first bridging of information flow control and anonymous communication to our knowledge. Additionally, we show that existing state-of-the-art protocols can be extended to support metadata privacy, by designing a novel protocol for deniable instant messaging (DenIM), which is a variant of the Signal protocol. To show the efficacy of our approach, we implement and evaluate a proof-of-concept instant messaging system running DenIM on top of unmodified Signal. We empirically show that the DenIM on Signal can maintain low-latency for unmodified Signal traffic without breaking existing features, while at the same time supporting deniable Signal traffic.

Traffic analysis attacks remain a significant problem for online security. Communication between nodes can be observed by network level attackers as it inherently takes place in the open. Despite online services increasingly using encrypted traffic, the shape of the traffic is not hidden. To prevent traffic analysis, the shape of a system's traffic must be independent of secrets. We investigate adapting the data-oblivious approach the reactive setting and present OblivIO, a secure language for writing reactive programs driven by network events. Our approach pads with dummy messages to hide which program sends are genuinely executed. We use an information-flow type system to provably enforce timing-sensitive noninterference. The type system is extended with potentials to bound the overhead in traffic introduced by our approach. We address challenges that arise from joining data-oblivious and reactive programming and demonstrate the feasibility of our resulting language by developing an interpreter that implements security critical operations as constant-time algorithms.

Thin hypervisors make it possible to isolate key security components like keychains, fingerprint readers, and digital wallets from the easily-compromised operating system. To work together, virtual machines running on top of the hypervisor can make hypercalls to the hypervisor to share pages between each other in a controlled way. However, the design of such hypercall ABIs remains a delicate balancing task between conflicting needs for expressivity, performance, and security. In particular, it raises the question of what makes the specification of a hypervisor, and of its hypercall ABIs, good enough for the virtual machines. In this paper, we validate the expressivity and security of the design of the hypercall ABIs of Arm's FF-A. We formalise a substantial fragment of FF-A as a machine with a simplified ISA in which hypercalls are steps of the machine. We then develop VMSL, a novel separation logic, which we prove sound with respect to the machine execution model, and use it to reason modularly about virtual machines which communicate through the hypercall ABIs, demonstrating the hypercall ABIs' expressivity. Moreover, we use the logic to prove robust safety of communicating virtual machines, that is, the guarantee that even if some of the virtual machines are compromised and execute unknown code, they cannot break the safety properties of other virtual machines running known code. This demonstrates the intended security guarantees of the hypercall ABIs. All the results in the paper have been formalised in Coq using the Iris framework.

Traffic analysis attacks pose a major risk for online security. Distinctive patterns in communication act as fingerprints, enabling adversaries to de-anonymise communicating parties or to infer sensitive information. Despite the attacks being known for decades, practical solution are scarce. Network layer countermeasures have relied on black box padding schemes that require significant overheads in latency and bandwidth to mitigate the attacks, without fundamentally preventing them, and the problem has received little attention in the language-based information flow literature. Language-based methods provide a strong foundation for fundamentally addressing security issues, but previous work has overwhelmingly assumed that interactive programs communicate over secure channels, where messages are undetectable by unprivileged adversaries. This assumption is too strong for online communication where packets can be trivially observed by eavesdropping. In this paper we introduce SELENE, a small language for principled, provably secure communication over channels where packets are publicly observable, and we demonstrate how our program level defence can reduce the latency and bandwidth overheads induced compared with program-agnostic defence mechanisms. We believe that our results constitute a step towards practical, secure online communication.

Nontransitive Noninterference (NTNI) and Nontransitive Types (NTT) are a new security condition and enforcement for policies which, in contrast to Denning's classical lattice model, assume no transitivity of the underlying flow relation. Nontransitive security policies are a natural fit for coarse-grained information-flow control where labels are specified at module rather than variable level of granularity. While the nontransitive and transitive policies pursue different goals and have different intuitions, this paper demonstrates that nontransitive noninterference can in fact be reduced to classical transitive noninterference. We develop a lattice encoding that establishes a precise relation between NTNI and classical noninterference. Our results make it possible to clearly position the new NTNI characterization with respect to the large body of work on noninterference. Further, we devise a lightweight program transformation that leverages standard flow-sensitive information-flow analyses to enforce nontransitive policies. We demonstrate several immediate benefits of our approach, both theoretical and practical. First, we improve the permissiveness over (while retaining the soundness of) the nonstandard NTT enforcement. Second, our results naturally generalize to a language with intermediate inputs and outputs. Finally, we demonstrate the practical benefits by utilizing state-of-the-art flow-sensitive tool JOANA to enforce nontransitive policies for Java programs.

Underneath every modern programming language is a runtime environment (RTE) that handles features such as automatic memory management and thread scheduling. In the information-flow control (IFC) literature, the RTE is often part of the trusted computing base (TCB), and there has been little focus on applying IFC to the implementation of the RTE itself. In this paper we address this problem by designing an IFC language, Zee, for implementing secure RTEs, thereby removing the RTE from the TCB. We implement Zee and design and implement secure versions of garbage collectors and thread schedulers using Zee. We also prove that a faithful calculus of Zee satisfies a strong variant of timing-sensitive noninterference.

This paper studies information flows via timing channels in the presence of automatic memory management. We construct a series of example attacks that illustrate that garbage collectors form a shared resource that can be used to reliably leak sensitive information at a rate of up to 1 byte/sec on a contemporary general-purpose computer. The created channel is also observable across a network connection in a datacenter-like setting. We subsequently present a design of automatic memory management that is provably resilient against such attacks.

Controlling confidential information in concurrent systems is difficult, due to covert channels resulting from interaction between threads. This problem is exacerbated if threads share resources at fine granularity. In this work, we propose a novel monitoring framework to enforce strong information security in concurrent programs. Our monitors are hybrid, combining dynamic and static program analysis to enforce security in a sound and rather precise fashion. In our framework, each thread is guarded by its own local monitor, and there is a single global monitor. We instantiate our monitoring framework to support rely-guarantee style reasoning about the use of shared resources, at the granularity of individual memory locations, and then specialize local monitors further to enforce flow-sensitive progress-sensitive information-flow control. Our local monitors exploit rely-guarantee-style reasoning about shared memory to achieve high precision. Soundness of rely-guarantee-style reasoning is guaranteed by all monitors cooperatively. The global monitor is invoked only when threads synchronize, and so does not needlessly restrict concurrency. We prove that our hybrid monitoring approach enforces a knowledge-based progress-sensitive non-interference security condition.

Information erasure is a formal security requirement that stipulates when sensitive data must be removed from computer systems. In a system that correctly enforces erasure requirements, an attacker who observes the system after sensitive data is required to have been erased cannot deduce anything about the data. Practical obstacles to enforcing information erasure include: (1) correctly determining which data requires erasure, and (2) reliably deleting potentially large volumes of data, despite untrustworthy storage services. In this paper, we present a novel formalization of language-based information erasure that supports cryptographic enforcement of erasure requirements: sensitive data is encrypted before storage, and upon erasure, only a relatively small set of decryption keys needs to be deleted. This cryptographic technique has been used by a number of systems that implement data deletion to allow the use of untrustworthy storage services. However, these systems provide no support to correctly determine which data requires erasure, nor have the formal semantic properties of these systems been explained or proven to hold. We address these shortcomings. Specifically, we study a programming language extended with primitives for public-key cryptography, and demonstrate how information-flow control mechanisms can automatically track data that requires erasure and provably enforce erasure requirements even when programs employ cryptographic techniques for erasure.

In capability-safe languages, components can access a resource only if they possess a capability for that resource. As a result, a programmer can prevent an untrusted component from accessing a sensitive resource by ensuring that the component never acquires the corresponding capability. In order to reason about which components may use a sensitive resource it is necessary to reason about how capabilities propagate through a system. This may be difficult, or, in the case of dynamically composed code, impossible to do before running the system. To counter this situation, we propose extensions to capability-safe languages that restrict the use of capabilities according to declarative policies. We introduce two independently useful semantic security policies to regulate capabilities and describe language-based mechanisms that enforce them. Access control policies restrict which components may use a capability and are enforced using higher-order contracts. Integrity policies restrict which components may influence (directly or indirectly) the use of a capability and are enforced using an information-flow type system. Finally, we describe how programmers can dynamically and soundly combine components that enforce access control or integrity policies with components that enforce different policies or even no policy at all.

Program progress (or termination) is a covert channel that may leak sensitive information. To control information leakage on this channel, semantic definitions of security should be progress sensitive and enforcement mechanisms should restrict the channel's capacity. However, most state-of-the-art language-based information-flow mechanisms are progress insensitive---allowing arbitrary information leakage through this channel---and current progress-sensitive enforcement techniques are overly restrictive.

In systems that handle confidential information, the security policy to enforce on information frequently changes: new users join the system, old users leave, and sensitivity of data changes over time. It is challenging, yet important, to specify what it means for such systems to be secure, and to gain assurance that a system is secure. We present a language-based model for specifying, reasoning about, and enforcing information security in systems that dynamically change the security policy. We specify security for such systems as a simple and intuitive extensional knowledge-based semantic condition: an attacker can only learn information in accordance with the current security policy. Importantly, the semantic condition is parameterized by the ability of the attacker. Learning is about change in knowledge, and an observation that allows one attacker to learn confidential information may provide a different attacker with no new information. A program that is secure against an attacker with perfect recall may not be secure against a more realistic, weaker, attacker. We introduce a compositional model of attackers that simplifies enforcement of security, and demonstrate that standard information-flow control mechanisms, such as security-type systems and information-flow monitors, can be easily adapted to enforce security for a broad and useful class of attackers.

We propose a new language-based approach to mitigating timing channels. In this language, well-typed programs provably leak only a bounded amount of information over time through external timing channels. By incorporating mechanisms for predictive mitigation of timing channels, this approach also permits a more expressive programming model. Timing channels arising from interaction with underlying hardware features such as instruction caches are controlled. Assumptions about the underlying hardware are explicitly formalized, supporting the design of hardware that efficiently controls timing channels. One such hardware design is modeled and used to show that timing channels can be controlled in some simple programs of real-world significance.

Mobile code is now a nearly inescapable component of modern computing, thanks to client-side code that runs within web browsers. The usual tension between security and functionality is particularly acute in a mobile-code setting, and current platforms disappoint on both dimensions. We introduce a new architecture for secure mobile code, with which developers can use, publish, and share mobile code securely across trust domains. This architecture enables new kinds of distributed applications, and makes it easier to reuse and evolve code from untrusted providers. The architecture gives mobile code considerable expressive power: it can securely access distributed, persistent, shared information from multiple trust domains, unlike web applications bound by the same-origin policy. The core of our approach is analyzing how flows of information within mobile code affect confidentiality and integrity. Because mobile code is untrusted, this analysis requires novel constraints on information flow and authority. We show that these constraints offer principled enforcement of strong security while avoiding the limitations of current mobile-code security mechanisms. We evaluate our approach by demonstrating a variety of mobile-code applications, showing that new functionality can be offered along with strong security.

Timing channels remain a difficult and important problem for information security. Recent work introduced predictive mitigation, a new way to mitigating leakage through timing channels; this mechanism works by predicting timing from past behavior, and then enforcing the predictions. This paper generalizes predictive mitigation to a larger and important class of systems: systems that receive input requests from multiple clients and deliver responses. The new insight is that timing predictions may be a function of any public information, rather than being a function simply of output events. Based on this insight, a more general mechanism and theory of predictive mitigation becomes possible. The result is that bounds on timing leakage can be tightened, achieving asymptotically logarithmic leakage under reasonable assumptions. By applying it to web applications, the generalized predictive mitigation mechanism is shown to be effective in practice.

We investigate techniques for general black-box mitigation of timing channels. The source of events is wrapped by a timing mitigator that delays output events so that they contain only a bounded amount of information. We introduce a general class of timing mitigators that can achieve any given bound on timing channel leakage, with a tradeoff in system performance. We show these mitigators compose well with other mechanisms for information flow control, and demonstrate they are effective against some known timing attacks.

A web mashup is a web application that integrates content from different providers to create a new service, not offered by the content providers. As mashups grow in popularity, the problem of securing information flow between mashup components becomes increasingly important. This paper presents a security lattice-based approach to mashup security, where the origins of the different components of the mashup are used as levels in the security lattice. Declassification allows controlled information release between the components. We formalize a notion of composite delimited release policy and provide considerations for practical (static as well as runtime) enforcement of mashup information-flow security policies in a web browser.

This paper studies the problem of securing information release in dynamic languages. We propose (i) an intuitive framework for information-release policies expressing both what can be released by an application and where in the code this release may take place and (ii) tight and modular enforcement by hybrid mechanisms that combine monitoring with on-the-fly static analysis for a language with dynamic code evaluation and communication primitives. The policy framework and enforcement mechanisms support both termination-sensitive and insensitive security policies.

Information security has a challenge to address: enabling information-flow controls with expressive information release (or declassification) policies. Existing approaches tend to address some aspects of information release, exposing the other aspects for possible attacks. It is striking that these approaches fall into two mostly separate categories: revelation-based (as in information purchase, aggregate computation, moves in a game, etc.) and encryption-based declassification (as in sending encrypted secrets over an untrusted network, storing passwords, etc.). This paper introduces gradual release, a policy that unifies declassification, encryption, and key release policies. We model an attacker's knowledge by the sets of possible secret inputs as functions of publicly observable outputs. The essence of gradual release is that this knowledge must remain constant between releases. Gradual release turns out to be a powerful foundation for release policies, which we demonstrate by formally connecting revelation-based and encryption-based declassification. Furthermore, we show that gradual release can be provably enforced by security types and effects.