Rate This Document
Findability
Accuracy
Completeness
Readability

Change History

Date

Description

2026-09-30

This issue is the sixth official release.

  • Released Ultrascan 5.8.0. Added mcsheng algorithm performance optimization and feedback-driven optimization for regular expression matching. Optimized the Grey configuration method to support setting and resetting process-level Grey configuration through public APIs.
  • Renamed KDADK-MMPM to Megascan, and released Megascan 1.0.0. Provided handler-compatible APIs as well as database, scratch, and stream scanning APIs, supporting rule logic combination, cross-segment stream scanning, and dynamic rule deletion at runtime.
  • Released KADD 1.1.0.
    • Updated the SQLi feature. Added the DL (CNN) inference mode to the SQLi detection library; added support for switching between ML and DL modes to the SQLi engine; adopted a unified policy of native SVE binaries (prioritized) + automatic fallback to ONNX Runtime for both the ML and DL engines; added the unified ONNX inference engine; optimized the configuration file structure to a flat format.
    • Added and released the PI feature. Provided a hybrid detection solution based on rules and deep learning (LERT model) to enable high-performance detection of jailbreak text.

2026-06-30

This issue is the fifth official release.

  • Renamed Hyperscan to Ultrascan, and released Ultrascan 5.7.0. Added the universal bytecode technology feature.
  • Renamed KDADK-APPID to KTAG, and released KTAG 1.0.0.
  • Released KADD 1.0.0. Added the SQLi detection library, which provides the lexical analyzer, feature extraction engine, and inference engine.

2026-03-30

This issue is the fourth official release.

  • Released Hyperscan 2.6.0, and optimized the Hyperscan short-byte (2–4 bytes) rule matching algorithm based on the new Kunpeng 920 processor model. For details, see Ultrascan.
  • Released KDADK-TOOLS 1.1.0, updated the header files and software packages of the feature extraction and inference APIs, and provided a reference example for network flow classification of Kunpeng BoostKit for Intelligent Data Analytics. For details, see KTAG.

2025-12-30

This issue is the third official release.

Reconstructed the "Hyperscan" section and added the description of the false-positive blocking technology.

2025-09-30

This issue is the second official release.

Changed the scenario name to Intelligent Data Analytics (IDA) and renamed DADK to KDADK.

2025-06-30

This issue is the first official release.