Authentication and Confidentiality

• Intelligent Service Security Application Protocol: Combines cross-platform communications with encryption, signature, and authentication, to improve IoT apps development capabilities.

• Fully implemented two-way authentication security scheme for IoT,
• Datagram Transport Layer Security (DTLS) protocol, based on RSA and designed for IPv6 over Low power Wireless Personal Area Networks (6LoWPANs), placed between transport and app player. It provides message integrity, confidentiality, and authenticity.

• Key Management System (KMS) has weaknesses of four major categories: key pool framework, mathematical framework, negotiation framework, and public key framework.
• The combinatorics-based KMS protocols suffer both connectivity and scalability, authentication.

Authentication and Confidentiality

• Most of the KMS protocols are not suitable for IoT. In fact, key pool ones suffer insufficient connectivity
• Key pool ones suffer insufficient connectivity
• Mathematical ones make use of the deployment knowledge to optimize the construction of their data structures, but such an approach cannot be used in IoT since client and server nodes are usually located in different physical locations; combinatorics-based KMS protocols suffer both connectivity and scalability/authentication

• Negotiation ones make use of the wireless channel and its inherent features to negotiate a common key, however they cannot be suitable for IoT because client and server nodes usually belong to different networks and they should route the information through the Internet in order to be able to talk with each other

Authentication and Confidentiality

• BLOM and Polynomial Schema is more suitable.
• Their computational overhead is quite low in comparison to a Public Key Cryptography (PKC) operations

•  Are the WSN proposals adaptable to the IoT environment, considering both the heterogeneity of the involved devices and the different application contexts?
• How and at which network layer to handle authentication? Is it feasible to reuse the traditional security mechanisms (e.g., encryption algorithms) or it is better to start from new solutions?
• How to handle the different keys?
• Which kind of key distribution mechanism is the most suitable?
• How to ensure an end-to-end integrity verification mechanism in order to make the system more resilient to malicious attacks?

Authentication and Confidentiality

• Suitable KMS protocols for IoT scenarios:  Blom and the polynomial schema.
• In such those schemes, several countermeasures are required to manage authentication and MitM attacks. Also presented a framework for IoT based on Public Key Infrastructure (PKI).

• Transmission model with signature-encryption schemes:
• Anonymity, Trustworthy and Attack Resistance) by Object Naming Service (ONS) queries.
• It provides identities authentication, platform creditability, data integrity.

Authentication and Confidentiality

• Also presented an authentication protocol using lightweight encryption based on XOR manipulation for anti-counterfeiting and privacy protection, coped with constrain IoT devices.

• And proposed an user authentication and key agreement scheme for WSN, by using hash and XOR computations. It ensures mutual authentication among users, sensor nodes and gateway nodes (GWN).

• Authentication and access control method, establishes session key on a lightweight encryption mechanism, Elliptic Curve Cryptography (ECC). This scheme defines attribute based access control policies, managed by an attribute authority, to enhance authentication.

Authentication and Confidentiality

• However, a clear solution to guarantee the confidentiality is still missing, and some efforts have been conducted in the WNS field

• Are the WSN proposals adaptable to the IoT environment, considering both the heterogeneity of the involved devices and the different application contexts?
• How and at which network layer to handle authentication? Is it feasible to reuse the traditional security mechanisms (e.g., encryption algorithms) or it is better to start from new solutions?
• How to handle the different keys?
• Which kind of key distribution mechanism is the most suitable?
• How to ensure an end-to-end integrity verification mechanism in order to make the system more resilient to malicious attacks?

Access Control

• Identified two subjects: data holders - feed data collectors with a specific target, and data collectors - identify and authenticate users and things from which informations are collected

• Layer responsible for data acquisition, presented a hierarchical access control scheme for this layer.
• It provides a single key and necessary keys by using a deterministic key derivation algorithm, for increasing the security and reducing nodes storage costs.

• Also presented an identity based system for personal location in emergency situations. It consists of: registration, users authentication, policy, and client subsystems.

• Represented streams as linear algebraic queries, provides the product authentication, by using the hash operations, modular additions/ multiplications and cryptographic security functions

Access Control

• In IoT we have also to deal with processing of streaming data and not, as in traditional database systems, with discrete data.
• The main critical issues in this context refer to performance and temporal constraints, since access control for a data stream is more computational intensive than in traditional DBMS
• Queries have to be directly executed on incoming streams, which can be made of large volumes of data that might arrive at unpredictable rates
• Layer responsibility for data acquisition, which is the direct responsible for the information collection. In such a layer, a large amount of nodes are required to sense a wide range of different data types for authorized users in accordance with privacy and security levels.

Access Control

• Proposed a stream-centric approach, which security constraints are directly embedded into data stream, reduces overhead, and enriches data streams with metadata called streaming tags

• Security constraints are directly embedded into data streams and not stored on the DSMS server.
• Security metadata tuples are interleaved with the data tuples in the streams, in order to reduce the overhead

• No new access control model is defined, but an enforcement mechanism suitable for streaming data, exploiting query processing

Access Control

• Major challenges related to access control :
• How to guarantee the access permission in an environment where not only users, but also things could be authorized to interact with the system?
• It is more effective to exploit a centralized or distributed approach or a semi-distributed one in order to manage the scalable IoT architecture?
• How to handle the huge amount of transmitted data (i.e., in the form of stream data) in a common recognized representation? How to support the identification of entities?

Privacy

​Any sensitive data for users, health issues, etc

• Data tagging for managing privacy in IoT is proposed.
• Using techniques taken from the Information Flow Control, data representing network events can be tagged with several privacy properties; such tags allow the system to reason about the flows of data and preserve the privacy of individuals.
• Although exploiting tagging within resource-constrained sensor nodes may not be a viable solution because tags may be too large with respect to the data size and sensitivity, therefore they generate an excessive overhead.

• A user-controlled privacy-preserved access control protocol is proposed, based on context-aware anonymity privacy policies.
• Privacy protection mechanisms are investigated: users can control which of their personal data is being collected and accessed, who is collecting and accessing such data, and when this happens.

Privacy

• Traditional privacy mechanisms are divided into two categories:
• Discretionary Access and Limited Access. The former addresses the minimum privacy risks, in order to prevent the disclosure or the cloning of sensitive data; whereas the latter aims at limiting the security access to avoid malicious unauthorized attacks

• Privacy risk with assignment of a static domain name to a specified IoT node.
• A privacy protection enhanced DNS (Domain Name System) for smart devices, which can authenticate the original users identity and reject illegal access to the smart device. The scheme is compatible with widely used DNS and DNSSEC (Domain Name System Security Extensions) protocols

Trust

Complex notion about which no definitive consensus exists in the scientific literature, although its importance is widely recognized.

• Main problem with many approaches towards trust definition is that they do not lend themselves to the establishment of metrics and evaluation methodologies
• Satisfaction of trust requirements are strictly related to the identity management and access control issues
• Trust Level Assessment of IoT entities: most smart objects are human-carried or human-related devices, so they are often exposed to public areas and communicate through wireless, hence vulnerable to malicious attacks.
•  Malicious nodes aim at breaking the basic functionality of IoT by means of trust related attacks: self-promoting, bad-mouthing and good-mouthing

Trust

• The trust management protocol for IoT : distributed, encounter-based, and activity-based: two nodes that come in touch to each other or involved in a mutual interaction can directly rate each other and exchange trust evaluation about the other nodes, so they perform an indirect rate which seems like a recommendation
• Therefore such a dynamic trust management protocol is capable of adaptively adjusting the best trust parameter setting in response to dynamically changing environments in order to maximize application performance.