A performance bug at a domestic Xinjiang test site (25-70Gbps) traced packet-processing lock contention to a plugin named "tsg_vulpes" calling an ONNX Runtime model for real-time "encrypted voice recognition" on live traffic; disabling this ML classifier resolved packet loss, indicating it's an optional, performance-costly add-on.
这一锁由tsg_vulpes使用，通过tsg-os-cli关闭该功能后，不再丢包运行正常...临时解决方案，关闭加密语音识别功能...set template name tsg_traffic_engine_default encrypt_traffic_identify voice_bahavior_engine no
Defense implications
- Geedge runs an ML classifier (ONNX-based) identifying voice traffic inside encrypted flows via behavioral features, not just signatures — voice-adjacent transport obfuscation should account for behavioral ML, not only fingerprint-based detection.
- This capability is toggleable with a known performance cost at high throughput — can't assume it's universally enabled in production.
Related findings
A dedicated internal "运营商前端分析团队" (Carrier Front-end Analysis Team) produces weekly, carrier-specific traffic-analysis reports explicitly titled "新疆移动流量分析报告" (Xinjiang Mobile Traffic Analysis Report) and "新疆联通流量分析报告" (Xinjiang Unicom Traffic Analysis Report), with instances dated from at least July 2022 through March 2024. One instance (Xinjiang Mobile, 2023-05-15) reports total carrier bandwidth of 4.4Tbps (1Tbps access bandwidth), 27.34PB of weekly throughput, 123 distinct identified application-layer protocols, and a section (4.1, pages 17+) explicitly tracking server IP/location for 28 named overseas apps (BBC, Discord, ESPN, Facebook, Gmail, Google, Hulu, Instagram, Netflix, Reddit, Skype, Telegram, Twitter, Uber, WhatsApp, Wikipedia, YouTube, Zoom, etc.).
A weekly "新疆移动流量分析报告" (Xinjiang Mobile Traffic Analysis Report), authored by the "运营商前端分析团队" (Carrier Front-End Analysis Team), directly ties the SNI/Server-IP overseas- APP report format to the China Mobile Xinjiang branch specifically, and states the pipeline identifies 126 distinct application-layer protocols including multi-layer tunnel nesting such as STUN.DTLS and STUN.RTP.RTCP.DTLS — i.e. it decomposes and classifies nested WebRTC-style transport stacks, not just top-level TLS/QUIC.
Multiple domestic IDC traffic-monitoring reports rank "Freegate" (a well-known GFW-circumvention tool) as a distinct, named top-10 application by traffic volume alongside Bytedance/Tencent/Alibaba/Baidu, with peak 18.92 Gbps and average 5.9 Gbps in one 2022-07-07 report -- confirming Freegate has its own dedicated app-ID classifier tracked at carrier-comparable traffic scale, not merely a low-volume/rare signature.
MESA Lab engineering logs (2020) describe development and TSG-environment testing of a SAPP plugin that identifies encrypted audio/video traffic and distinguishes on-demand from live streams. The workflow uses Fiddler (a TLS MITM proxy) to decrypt traffic client-side and auto-label training data for the underlying ML model, and required a dlopen() workaround to load libpython.so for the model at runtime.
An internal TSG operations/troubleshooting manual lays out TSG's full traffic pipeline (NIC -> mrzcpd/marsio capture driver -> sapp DPI engine -> firewall/proxy(KNI->TFE)/active-defense/WAN-NAT policy branches) and shows engineers using maat_redis_tool to pull the live Redis-synced blocking policy tables (TSG_SECURITY_COMPILE, TSG_OBJ_IP_ADDR, TSG_OBJ_APP_ID) and filter them by numeric policy/object ID to debug why a block rule isn't firing.
The internal 'MAAT网络流处理配置统一描述框架' engineering manual (v3.1.20, author 郑超, 2021) documents MAAT's Redis-synced rule-compilation framework and its RuleScan/Hyperscan-based pattern-matching engine (libmaatframe.so / librulescan), and records that RuleScan's fast-scan feature caused a production outage on 2019-03-19 and has been disabled ever since.