The new decentralized computing paradigm introduced by Machine-to-Machine (M2M) communications and the Internet of Things (IoT) ecosystem requires developing new security mechanisms and frameworks, adapted to this new architecture. The variety of IoT use cases includes applications leveraging low-level devices such as sensor or actuators, to applications deploying safety critical devices such as connected vehicles in Intelligent Transportation Systems (ITS). Devices are deployed as nodes in communication networks, and have become in recent years targets for attackers who exploit the resource-constrained nature of the devices in order to compromise the safety, security, and availability of the different applications. Two of the main challenges in this ecosystem are securing the communication between IoT devices, and ensuring that devices in the network have not been compromised or tampered with, thus attesting of the integrity of the entire network. The challenges are exacerbated by the nature of devices, which present stringent constraints, notably in terms of computational capabilities, storage space, and energy resource. In addition, new privacy concerns affecting users in IoT applications have risen, and require implementing privacy-friendly authentication and attestation mechanisms.Authentication mechanisms allow systems to identify themselves on the network, and provide solutions for the first challenge. Remote Attestation is a security mechanism which enables control systems to verify the software state of devices in the network, thus detecting any tampering or remote malware injection attacks.In this thesis, we aim to contribute to the development of new and privacy-preserving authentication and attestation mechanisms, which are particularly adapted for implementation in constrained environments.In the first part of this thesis, we leverage a cryptographic mechanism deployed in trusted computing, namely Direct Anonymous Attestation (DAA), in order to provide decentralized, and privacy-preserving authentication protocols adapted for constrained environments. Our work contributes to the development of a variant of Direct Anonymous Attestation schemes, called pre-Direct Anonymous Attestation (pre-DAA), which achieves a trade-off between security andefficiency that was not previously achieved in the literature. In particular, our pre-DAA scheme is proven secure in the Random Oracle Model (ROM) under the q-Strong Diffie Hellman (q—SDH) assumption, while performing better than DAA schemes proven secure under an interactive assumption. The pre-DAA scheme is subsequently used in the development of two privacy-preserving authentication protocols. The first application of our pre-DAA scheme consists in the design of a decentralized architecture for secure communication in vehicular ad hoc networks, which removes the need for a centralize Public Key Infrastructure. The second application of our pre-DAA scheme is the design of a mobile-based access control protocol for public transport systems, which addresses the issue of user traceability inherent to current access control control protocols for transport systems.In the second part of this thesis, we address the device integrity verification challenge by designing a remote attestation protocol which enables the secure and efficient attestation of groups (or swarms) of devices. Our attestation protocol verifies the integrity of every device in the network during a single attestation phase, by leveraging the aggregating properties of an aggregate algebraic MAC scheme. Compared to swarm attestation protocols in the literature, our contribution enables the detection of an erroneous attestation report in the aggregated result, thus allowing the identification of compromised devices.