Industrial Internet Security Framework v 1.0 | Page 17

Security Framework 3: Key System Characteristics Enabling Trustworthiness the physical processes of the system. Data integrity, a subset of integrity, ensures that unauthorized parties cannot alter data and take control of the system without detection. Availability is the property of on-demand, timely and reliable access to and use of information by an authorized user. The systems responsible for controlling the physical process should provide continuous control and oversight by human operators of the physical process. A human may need to intervene in the case of an attack, for example to shut the system down. Availability controls generally involve redundancy and engineering change control. Sometimes they include security activities that find and mitigate software vulnerabilities that create unreliable execution, visualization or resource consumption that negatively affect the systems. In traditional operational technology (OT) systems, availability has been considered paramount, followed by integrity, with confidentiality generally being the last consideration, leading to the acronym AIC (also known as the security triad). 3.3 SAFETY Safety is the condition of the system operating without causing unacceptable risk of physical injury or damage to the health of people, either directly or indirectly, as a result of damage to property or to the environment. Assurance of safety endeavors to eliminate both systematic and probabilistic failures. Traditional OT safety-assessment techniques focus on physical items and processes, then combine empirically derived component failure probabilities into total system risk. Risk analysis to identify hazards intends to prevent faulty operations and improve system resilience to unexpected events. However, a software component always behaves exactly as it is programmed; it is not possible to make useful statistical characterizations of software failures. If a software component has never misbehaved during testing, it may not have been exposed to a sequence of inputs that would have uncovered the defect. Test coverage does not necessarily correlate to failure rate. Approaches for managing probabilistic failures do not address threats because adversaries will be able to exploit security-related systematic failures reliably once those vulnerabilities have been discovered. Traditional efforts for industrial software focused on functional correctness and did not assume that an adversary was involved. In today’s connected systems, a remote attacker is able to exploit weaknesses 1 to drive the system into an unsafe state. This contrasts sharply with traditional IT security, where a security analysis of the threat and threat-actor skills and capabilities is used to determine the likelihood of weaknesses that can be exploited. Many of the same tools, techniques and practices used to produce safety-critical software can also identify, remove and mitigate potential security weaknesses. Many safety regulations and 1 For a public collection of weaknesses, see [CWE] IIC:PUB:G4:V1.0:PB:20160926 - 17 -