13-SW Dependability and Reliability

Dependability Importance

• Critical property of computer-based systems.
• Reflects user's trust in system operation and failure prevention.
• Attributes: reliability, availability, safety, security, resilience.
• Failures can have widespread negative impacts.
• Rejection by users if system is undependable.
• High costs associated with failures (economic loss, damage).

Dependability Properties

• Availability: Probability system is operational and provides services.
• Reliability: Probability of correct service delivery as expected.
• Security: Ability to resist unauthorized or accidental intrusions.
• Safety: Maintaining service continuity despite disruptive events (e.g., failures).
• Resilience: Ability to recover from damages.

Attribute Dependencies

• Safe operation relies on availability and reliability.
• Unreliability can stem from data corruption or attacks.

Dependability Development

• Avoid introduction of errors, effective verification and validation (V&V).
• Design for fault tolerance and protection mechanisms.
• Configure systems appropriately for their environment.

Cost of Dependability

• Costs increase exponentially with higher dependability requirements.
• Use of expensive techniques and extensive testing drives costs.

Economics of Dependability

• Decision between trustworthiness and cost of failure.
• Fault tolerance essential in critical systems (e.g., aviation).

Fault-tolerant Architectures

• Based on redundancy and diversity to ensure reliability.
• Redundant components provide backups; diverse components mitigate common failures.

Complexity and Redundancy

• Increasing redundancy/completeness complicates system design, increasing error potential.
• Simplicity advocated by some engineers for better dependability outcomes.

Verification and Validation Process

• Utilize diverse approaches to prevent process errors.
• Explicit, repeatable, and auditable processes are crucial.

Dependable Programming Guidelines

1. Limit information visibility.
2. Validate inputs, including bounds.
3. Provide exception handling.
4. Avoid complex code structures.

Faults and Errors

• Clear definitions between human errors, system faults, system errors, and failures outlined.
• Not all faults result in errors; not all errors cause failures.

Reliability

• Defined as the probability of failure-free operation over time.
• Perceived reliability can diverge from actual reliability.

Availability

• Measures operational probability at a given time.
• Perception of availability varies based on impact and duration of service disruptions.

Reliability Measurement

• Metrics are crucial for assessing system reliability.
• Include probabilities of failure on demand and mean time to failure.

Specification of Reliability

• Necessary to clarify stakeholder needs and assess testing requirements.
• Challenges in developing appropriate operational profiles and test data.

Achieving Dependability

• Strategies include avoidance, resistance, detection, recognition, and recovery for enhancing dependability and security.

Conclusion

• Focus on creating dependability in system architecture, programming, and validation processes to ensure high reliability and availability in critical systems.