Lightz-AIO | Implement rules splitting based on alignment stats

Objective

Implement an intelligent rule-splitting mechanism based on execution times calculated during the alignment scan. This aims to optimize multi-core scanning for balanced execution time.

Background

This issue is part of the larger effort to "Optimize multi-core scanning for balanced execution time" in GLAS.

Proposed Solution: Distribution Algorithm

We will implement a greedy distribution algorithm with the following steps:

1. Cache Loading:
   • Load the execution time cache at startup.
   • This cache should contain rule execution times from previous alignment scans.
2. Rule Sorting:
   • Sort rules by their execution time, from longest to shortest.
3. Greedy Distribution:
   • Start with the longest-running rules.
   • Assign each rule to the core with the lowest total assigned time.
   • This approach aims to balance the workload across all available cores.
4. Fallback Mechanism:
   • If no cache is available, fall back to the current method of rule distribution.

Verification

• The MR related to the algorithm improvement has been verified here.
• GLAS has been released here.