Fault-Tolerant System Design

Explore fault-tolerant system design principles and best practices for robust application development in this conference talk from the International PHP Conference. Dive into the internals of a high-load mobile application with 13 million users worldwide, processing over a million transactions daily. Learn about multi-fault tolerant systems, common mistakes to avoid, and techniques for handling errors and exceptions. Discover strategies for software refactoring, performance optimization, and reliability models. Gain insights on fault handling techniques, failure response stages, and data separation. Understand the importance of separation of concerns, exception handling, and edge case management in creating resilient software systems.

Intro
URAL Telekom . Secure Communication software . Software Refactoring for Testability Performance optimization
Fault-tolerant System design • Robust Software Development Tools and techniques
Fault Handling Techniques . Fault Avoidance • Fault Detection • Masking Redundancy • Dynamic Redundancy
Failure Response Stages . Fault detection and Diagnosis • Fault isolation • Reconfiguration • Recovery
Reliability Models . Serial Parallel
Reconfigure . Use redundant system Graceful degradation • Indicate degraded state
Data separation . Separate Metadata from data Separate control from workload
Reliability . Can be accomplished using redundancy Except for design faults
Software faults are mostly . Software specifications • Design error • Developer error • Unexpected conditions
Separation of Concerns • Split code into modules • No direct data access • No direct data modification! • Update data through a dedicated Repository or Service
Exception handling • Handle unknown and unpredictable faults Adds to Fault tolerance • Decide where to catch those exceptions
Error recovery • Backward recovery Forward recovery
Edge case handling . Code review