Background Explanation

As the number of users grows and functional demands increase, we have observed that some high-resource consumption configuration options may lead to service instability. To ensure service quality, we conducted a thorough analysis and developed corresponding optimization strategies.

Resource Optimization Strategies

1. Filter Update Mechanism Optimization

Current Situation Analysis

• Some users have set filters to update hourly
• Each update requires a complete download-parse-deduplication process
• International bandwidth limitations cause longer update times
• Servers experience sustained high loads

Optimization Plan

We will adjust the update interval to a minimum of 72 hours for the following reasons:

• Most filter lists have update cycles of 24-72 hours
• Reduce unnecessary resource consumption
• Ensure service stability
• Optimize bandwidth usage efficiency

Impact Assessment

• Positive Impact
  • More stable service response
  • More reasonable resource usage
  • Reduced system load
• Minimal Impact
  • Rule updates remain within a reasonable cycle
  • No impact on protection effectiveness

2. Parallel Request Strategy

Current Situation

Currently, most users have enabled the parallel request feature, but the benefits are limited under the existing architecture:

• The latency difference of upstream services on Alibaba Cloud is usually within 5ms
• May trigger request frequency limits of Alibaba Cloud’s public services
• Increases unnecessary system overhead

Usage Recommendations

• Recommend using load balancing mode
• Parallel requests are suitable for the following scenarios:
  • Significant latency differences in upstream services (>200ms)
  • Situations where service quality is unstable
  • Cross-border access scenarios

Note: No throttling issues due to parallel requests have been found so far; this feature remains open for now.

3. Third-Party List Management

Safety Considerations

To ensure system stability, we have temporarily disabled support for some third-party lists:

• The scale of external lists is unpredictable
• May lead to resource overload
• Service stability cannot be guaranteed

Future Plans

We are researching safer third-party list management solutions to potentially reopen this feature in the future.