Design Principles

The following are common principles for optimizing performance:

• Centralization: Identify key workloads, reduce unnecessary processing, and enhance the most important parts for better performance.
• Localization: Select the resources close to activities, functions, and results, and avoid indirect approaches that may increase traffic or workload and harm performance.
• Resource sharing: Share resources and reduce the contention for resources to improve the overall performance. For example, multiple processes can read the same part of a database simultaneously.
• Parallel processing: Leverage parallel processing to make up for the communication overhead and resource contention delay.
• Load distribution: Handle loads at different time or locations. Divide resources into small separate resource groups and allow different processes or threads to access them independently. Schedule multiple concurrent requests for processing within a time segment.