Abstract
Adding new telemetry to mobile operating systems (OS) requires changing code and pushing updated binaries to field devices, which can have a substantial lag before new telemetry reaches an adequate number of field devices. This hampers the optimization and debugging of operating systems and the apps they host. This disclosure describes a dynamic instrumentation framework that enables rapid pushing of new telemetry configurations to field devices. The configurations can identify parts of the mobile OS (and apps running thereon) to be inspected and specifications on how to extract information from the identified parts. In contrast to existing techniques of disseminating new binaries, configurations pushed to devices enable telemetry collection without requiring changes to the kernel or user-space code running on the device. The resulting mechanism retains comprehensive metrics-collection abilities without the need to push updated binaries. The mechanism can be used to rapidly collect newly defined metrics at scale. The results of telemetry can be sent to a server in a secure, private, and performant manner.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Tseng, Yu-Ting; Murray, Tim; Lee, Il-Sung; Gilbride, Matt; and Ke, Hao, "Dynamic Instrumentation Framework for Operating System Telemetry", Technical Disclosure Commons, (September 01, 2024)
https://www.tdcommons.org/dpubs_series/7324