Large numbers of smart connected devices, also named as the Internet of Things (IoT), are permeating our environments (homes, factories, cars, and also our body - with wearable devices) to collect data and act on the insight derived. Ensuring software integrity (including OS, apps, and configurations) on such smart devices is then essential to guarantee both privacy and safety. A key mechanism to protect the software integrity of these devices is remote attestation: A process that allows a remote verifier to validate the integrity of the software of a device. This process usually makes use of a signed hash value of the actual device's software, generated by dedicated hardware. While individual device attestation is a well-established technique, to date integrity verification of a very large number of devices remains an open problem, due to scalability issues. In this paper, we present SANA, the first secure and scalable protocol for efficient attestation of large sets of devices that works under realistic assumptions. SANA relies on a novel signature scheme to allow anyone to publicly verify a collective attestation in constant time and space, for virtually an unlimited number of devices. We substantially improve existing swarm attestation schemes by supporting a realistic trust model where: (1) only the targeted devices are required to implement attestation; (2) compromising any device does not harm others; and (3) all aggregators can be untrusted. We implemented SANA and demonstrated its efficiency on tiny sensor devices. Furthermore, we simulated SANA at large scale, to assess its scalability. Our results show that SANA can provide efficient attestation of networks of 1,000,000 devices, in only 2.5 seconds.

Large numbers of smart devices are permeating our environment to collect data and act on the insight derived. Examples of such devices include smart homes, factories, cars, or wearables. For privacy, security, and safety, ensuring correctness of the configuration of these devices is essential. One key mechanism to protect the software integrity of these devices is attestation.

Today, large numbers of smart interconnected devices provide safety and security critical services for energy grids, industrial control systems, gas and oil search robots, home/office automation, transportation, and critical infrastructure. These devices often operate in swarms -- large, dynamic, and self-organizing networks. Software integrity verification of device swarms is necessary to ensure their correct and safe operation as well as to protect them against attacks. However, current device attestation schemes assume a single prover device and do not scale to swarms. We present SEDA, the first attestation scheme for device swarms. We introduce a formal security model for swarm attestation and show security of our approach in this model. We demonstrate two proof-of-concept implementations based on two recent (remote) attestation architectures for embedded systems, including an Intel research platform. We assess performance of SEDA based on these implementations and simulations of large swarms. SEDA can efficiently attest swarms with dynamic and static topologies common in automotive, avionic, industrial control and critical infrastructures settings.

In recent years, Cloud Computing has gained remarkable popularity due to the economic and technical benefits provided by this new way of delivering computing resources. Businesses can offload their IT infrastructure into the cloud and benefit from rapid provisioning, scalability, and cost advantages. While cloud computing can be implemented on different abstraction levels, we focus on Infrastructure Clouds such as Amazon EC2 [1] that provide virtual machines, storage, and networks.

Virtual data centers allow the hosting of virtualized infrastructures (networks, storage, machines) that belong to several customers on the same physical infrastructure. Virtualization theoretically provides the capability for sharing the infrastructure among different customers. In reality, however, this is rarely (if ever) done because of security concerns. A major challenge in allaying such concerns is the enforcement of appropriate customer isolation as specified by high-level security policies. At the core of this challenge is the correct configuration of all shared resources on multiple machines to achieve this overall security objective. To address this challenge, this paper presents a security architecture for virtual data centers based on virtualization and Trusted Computing technologies. Our architecture aims at automating the instantiation of a virtual infrastructure while automatically deploying the corresponding security mechanisms. This deployment is driven by a global isolation policy, and thus guarantees overall customer isolation across all resources. We have implemented a prototype of the architecture based on the Xen hypervisor.

We address the problem of integrity management in a virtualized environment. We introduce a formal integrity model for managing the integrity of arbitrary aspects of a virtualized system. Based on the model, we describe an architecture called PEV, which stands for protection, enforcement, and verification. The architecture generalizes the integrity management functions of the Trusted Platform Module (TPM) to cover not just software binaries, but also VMs, virtual devices, and a wide range of security policies. The architecture enables the verification of security compliance and enforcement of security policies. We describe a prototype implementation of the architecture based on the Xen hypervisor. We demonstrate the policy enforcement and compliance checking capabilities of our prototype through multiple use cases.

We describe a secure network virtualization framework that helps realize the abstraction of Trusted Virtual Domains (TVDs), a security-enhanced variant of virtualized network zones. The framework allows groups of related virtual machines running on separate physical machines to be connected together as though there were on their own separate network fabric and, at the same time, helps enforce cross-group security requirements such as isolation, confidentiality, security, and information flow control. The framework uses existing network virtualization technologies, such as Ethernet encapsulation, VLAN tagging, and VPNs, and combines and orchestrates them appropriately to implement TVDs. Our framework aims at automating the instantiation and deployment of the appropriate security mechanism and network virtualization technologies based on an input security model that specifies the required level of isolation and permitted network flows. We have implemented a prototype of the framework based on the Xen hypervisor. Experimental evaluation of the prototype shows that the performance of our virtual networking extensions is comparable to that of the standard Xen configuration.

The application of formal methods for rigorously validating cryptographic protocols has been getting increasing attention. The de facto standard for modeling such protocols in formal proof systems is the Dolev-Yao model that, e.g., uses abstract encryption instead of cryptographic encryption primitives. The Dolev-Yao model has been originally intended and successfully used for detecting flaws in many protocols. However, recent publications claim to perform actual proofs of security using this model, i.e., absence of any attack. We doubt this claim and challenge Dolev-Yao-based models as being oversimplified for establishing security proofs against arbitrary attacks.We substantiate our claim by an example protocol. This protocol has been proven secure in a Dolev-Yao-based model using formal methods. In a later publication, the protocol has been broken by describing a cryptographic attack. The attack was not detected in the formal analysis since any Dolev-Yao-based model only comprises a predefined set of adversary capabilities. The particular attack to break the protocol was not comprised.The only reliable long-term remedy is to proof resilience against all attacks (both known and unknown ones). Recent approaches on cryptographic models of security have already made great progress towards this goal. Unfortunately, proofs in these are more complex and harder to automate. On the short run, it therefore is appropriate to improve the quality of formal analysis without striving for complete proofs. This can be achieved by means of evolving a catalog of adversary capabilities. Future formal analysis can then show resilience against any attack in this catalog. We initiate this discussion on an "adversary capability catalog" by providing a cryptographer's wish list. This list that points out several features which approaches based on the Dolev-Yao model or future extensions of it should cover in order to be effective for cryptographic protocol verification.

Privacy is an increasing concern in the marketplace. Although enterprises promise sound privacy practices to their customers, there is no technical mechanism to enforce them internally In this paper we describe a privacy policy model that protects personal data from privacy violations by means of enforcing enterprise-wide privacy, policies. By extending Jajodia et al.'s flexible authorization framework (FAF) with grantors and obligations, we create a privacy control language that includes user consent, obligations, and distributed administration. Conditions impose restrictions on the use of the collected data, such as modeling guardian consent and options. Access decisions are extended with obligations, which list a set of activities that must be executed together with the access request. Grantors allow to define a separation of duty between the security officer and the privacy officer.

Regulations and consumer backlash force many organizations to re-evaluate the way they manage private data. As a first step, they publish privacy promises as text or P3P. These promises are not backed up by privacy technology that enforces the promises throughout the enterprise. Privacy tools cover fractions of the problem while leaving the main challenge unanswered.This article describes a new approach towards enterprise-wide enforcement of the privacy promises. Its core is a new framework for managing collected personal data in a sensitive, trustworthy way. The framework enables enterprises to publish clear privacy promises, to collect and manage user preferences and consent, and to enforce the privacy promises throughout the enterprise.One of the foundations of this framework is the "sticky policy paradigm" that defines a customer centric model for managing policies, preferences, and consent.

A contract is a non-repudiable agreement on a given contract text, i.e., a contract can be used to prove agreement between its signatories to any verifier. A contract signing scheme is used to fairly compute a contract so that, even if one of the signatories misbehaves, either both or none of the signatories obtain a contract. Optimistic contract signing protocols use a third party to ensure fairness, but in such a way that the third party is not actively involved in the fault-less case. Since no satisfactory protocols without any third party exist, this seems to be the best one can hope for. We prove tight lower bounds on the message and round complexity of optimistic contract signing on synchronous and asynchronous networks, and present new and efficient protocols based on digital signatures which achieve provably optimal efficiency. 1 Introduction A contract is a non-repudiable agreement on a given text [4]. A contract signing scheme includes at least three players and two protocol...

This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents and will be distributed outside of IBM up to one year after the date indicated at the top of this page. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties).