Pattern-Based Configuration for Minimizing Recovery Impact Zones

The patent applies preliminary action by pre-defining configuration patterns that specify attribute values and recovery actions before the actual configuration generation. These patterns are stored in a data structure and used to guide the configuration generation process, allowing the system to select configurations based on predefined optimal patterns rather than exploring all possibilities from scratch. This reduces the search space and time required while maintaining high service availability.

The patent changes the approach from exploring all possible configuration parameters to selecting from predefined configuration patterns. Each pattern specifies particular attribute values (such as redundancy models, recovery actions, and impact zones) that have been optimized in advance. By transforming the configuration generation into a pattern matching and selection problem rather than a full combinatorial search, the system achieves both high reliability and reduced processing time.

The patent segments the configuration generation process into distinct phases: first identifying configuration patterns from predefined options, then matching patterns to requirements, and finally generating configurations based on selected patterns. This segmentation separates the complex combinatorial search from the actual configuration generation, reducing the complexity of the generation process itself while maintaining optimized service availability through the pre-defined patterns.

The patent introduces configuration patterns as an intermediary between the requirements and the final configurations. Instead of directly generating configurations through complex combinatorial exploration, the system uses patterns as intermediate representations that capture optimal configuration characteristics. This intermediary layer simplifies the generation process by providing a structured approach to selecting and combining configuration elements.

The patent applies preliminary action by pre-identifying and storing configuration patterns that satisfy all necessary criteria before the generation process begins. These patterns are prepared in advance with their associated attribute values and recovery actions, so that during configuration generation, the system only needs to match patterns to requirements rather than evaluating all possible configurations against multiple criteria. This eliminates premature pruning while maintaining high generation speed.

The patent changes the selection approach from evaluating configurations against multiple criteria during generation to matching predefined patterns that already encode the optimal parameter combinations. By transforming the problem into pattern matching, the system achieves both fast processing (no need to evaluate all combinations against multiple criteria) and high quality (patterns are pre-optimized to satisfy all requirements).