Software Patch LKG Cache for Binary Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing software patch systems are inefficient due to lengthy release times, quality risks, and unnecessary testing and download costs, as they include unnecessary binary changes and require thorough testing of all binaries in a patch, even if only a few are new.
Innovation Solution
Implementing a last known good (LKG) method that uses file caches for previously tested binaries, prioritizing the GDR cache over the SP cache, and synchronizing dependency databases to selectively include only necessary binaries in patches, reducing testing and deployment time and costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all binaries are included in the patch package, then completeness of the patch is improved, but the patch size and download cost increase
Solution Approach 1:
The patent extracts only the necessary binaries from the complete set of binaries. By comparing the dependency information with the binaries available in the LKG cache, the system identifies and includes only those binaries that are actually needed for the fix, excluding redundant ones from the patch package.
Solution Approach 2:
The patent segments the patch package into two parts: binaries that need to be updated (new or modified) and binaries that can remain at their last known good version. This segmentation allows the system to include only the necessary updated binaries in the patch while excluding binaries that don't need changes.
2Reliability
If all binaries are tested before release, then quality assurance is improved, but the testing time increases
Solution Approach 1:
The patent extracts only the binaries that actually need testing from the complete set of binaries in the patch. By using dependency information to identify which binaries are new or modified, the system tests only those specific binaries rather than all binaries in the package, significantly reducing testing time while maintaining quality assurance.
3Adaptability or versatility
If dependent binaries are included in the patch, then functionality is improved, but unnecessary changes and testing requirements increase
Solution Approach 1:
The patent extracts only the necessary dependent binaries from the complete set of dependent binaries. By comparing dependency requirements with the LKG cache contents, the system identifies which dependent binaries are already in a good state and can be excluded from the patch, including only those that actually need to be updated.
Data Source
AI summary
A system and method for last known good (LKG) for patches. A file cache is created to store binaries which have shipped as part of a previous wide public release—security fixes (GDRs) and service packs. There are two levels of cache, including an SP cache which keeps all of the files which were shipped as part of the service pack, and a GDR cache which keeps all of the files shipped in the GDRs. Customers do not need to download all of the binaries contained in a package when they already have the version of the dependent binaries on their system when they get the new patch. Customers only need to test the binaries which have changed, rather than everything in the new package, which speeds up the deployment of critical patches.