Assurance is a measure of confidence in a system based upon a composition of its trust, correctness, integrity, security, and reliability. Cyber-assurance is defined as a means of internet of things (IoT) smart devices and networks providing the opportunity of automatically securing themselves against security threats; the concept of cyber-assurance must provide embedded security within these IoT devices to allow these new networks to operate correctly even when subjected to a cyber-attack. Assurance is the evidence, which convinces us that an above-defined property holds. Techniques such as testing, disciplined development, formal methods, and others to build up evidence for each of these desired properties. This chapter defines trust as confidence based on the available evidentiary mechanisms that the software that will behave reliably and correctly while maintaining the integrity and security of itself and the system in which it is embedded. An assurance strategy is a plan for how to provide the evidence that a system merits our trust.
TopIntroduction
The internet of things (IoT) refers to the network or networks encompassing the use of standard Internet Protocol (IP) technologies to connect people, processes, and things to enable new cyber-physical systems (Brooks, 2020). Whereas previously, the Internet has generally been understood as comprising a network of computers, the adoption of mobile, embedded sensors and other technologies is expanding this definition to include people and objects outfitted or embedded with smart sensors (Brooks, 2017). As this trend grows over time with improved technology and less expensive hardware, the number of connected “objects” will trend toward all-encompassing devices, sensors, instrumentation, mobile and fixed assets and people. Cyber-assurance, the justified confidence that networked systems are adequately secure to meet operational needs under a cyber-attack, is required for IoT devices and networks (Brooks & Park, 2016).
Reactions to the cyber-assurance dilemma range across the spectrum from denial-refusal to use software components for high-assurance applications, to defeatism-abdication of quality control of any kind because it all seems too hard (Brooks, 2017). While the former is merely unrealistic, the latter is irresponsible. Edge-based systems for the IoT are here; it now falls to us to develop a cyber-assurance methodology to enable us to use them safely (Brooks, 2017). Fail-safe design, redundancy, tamper-detection were all once new ideas that, once embraced, radically changed both evaluation and development. A mature assurance model for Edge-based systems will require both new tools and new practices (Cao et al., 2020).
Edge computing constitutes a new concept in the computing landscape as the technology brings the service and utilities of cloud computing closer to the end user and is characterized by fast processing and quick application response time (Wang et al., 2020). Edge computing refers to locating applications – and the general-purpose compute, storage and associated switching and control functions needed to run them – relatively close to end users and/or IoT endpoints (Want et al., 2020). This greatly benefits applications performance and associated quality of experience (QoE) and it can also improve efficiency and thus the economics depending on the nature of the specific application. Edge computing is also important for localization of data and efficient data processing. Industry and Government regulations may require localization of data for security and privacy reasons. Certain application scenarios may pose restrictions on the use of excessive transport bandwidth or may require transport to external sites to be scheduled by time-of-day, requiring local storage or caching of information. Additionally, there may need to be local processing of information to reduce the volume of traffic over transport resources.
As the cyber mission has evolved to encompass the availability, integrity, authentication, confidentiality and non-repudiation requirements of information systems, a solution set has arisen to accomplish that mission which relies on elements of protection, detection and response. Our software trust strategy must be similarly multidimensional. While prevention is clearly the best cure, in practice, achieving the desired level of trust in a system containing software will require active components, mainly containment and detection, to compensate for what can’t engineered into protection (Weyns, 2020).