Software Fault Tolerance – Lecture Overview

Vocabulary flashcards covering key concepts from the software fault tolerance lecture.

Software fault tolerance

Methods and mechanisms that reduce the negative influence of bugs on availability, reliability, security, and correctness.

Availability

The ability of a service to be reachable by clients with maximum uptime (often aimed for no downtime).

Reliability

Output is correct and reproducible for the same input; results are dependable over time.

Security

Measures to prevent unauthorized access and protect data from leakage or tampering.

Correctness

The output or behavior of a service adheres to its formal specification.

Cloud services

Software running in shared cloud infrastructures (private/public/hybrid) rather than in a private data center.

Continuous delivery

Frequent, small releases with automated testing, rollout, and rollback to enable rapid feedback and updates.

Containers

Lightweight virtualization that packages software for easy shipping and consistent execution, well-suited for microservices.

Microservice architecture

An approach where a system is built from small, independent services that can be deployed, scaled, and updated separately.

Redundancy

Having multiple components or paths to maintain service availability during failures or maintenance.

Rollback

Automated process to revert to a previous stable release when a new release has issues.

Defect

A detectable flaw in code that can lead to bugs; measurable as defects per lines of code.

Bug

An error or flaw in a program that can cause incorrect or unexpected behavior.

Defect density

The number of defects per 1,000 lines of code, used as a quality metric.

Static analysis

Examining source code without executing it to identify potential issues or violations.

Dynamic analysis

Observing program behavior by executing the code to find defects during runtime.

Coverage report

A study that measures defects across large code bases, highlighting defect density and outstanding issues.

Byzantine fault tolerance

A failure model where components may behave arbitrarily or maliciously, posing a high challenge to tolerate.

Crash failure

A failure where a program stops running and its state is lost.

Data corruption

An error state where data is altered incorrectly, not always visible immediately to users.

Error propagation

Propagation of an error from one component to others through data exchange, causing cascading failures.

Fault

A defect or flaw in a system that can activate an error.

Error

A system state that deviates from the specification, caused by a fault.

Failure model

A defined set of faults that a system aims to tolerate, setting the scope of fault tolerance design.

Hardening interfaces

Designing interfaces to be robust against faults to prevent error propagation.

Rejuvenation

A fault tolerance mechanism that periodically refreshes components to recover from leaks or degradation.

Rx (Recovery and Re-execute)

A pattern that helps a program continue by recovering or reconfiguring the environment after faults.

Orchestra

A fault-tolerance concept focusing on system call interfaces and synchronization to improve robustness.

Buxton's deviant behavior

A static-analysis approach to detect contradictions or deviant behavior in code.

Taunt analysis

Taint analysis: tracking data flow to identify security flaws and vulnerabilities by marking tainted data.