-
Notifications
You must be signed in to change notification settings - Fork 2
MobileInsight
MobileInsight (short for MI, hereafter) is an on-phone tool to gain open access to operational cellular network operation information which remained closed before. It allows researcher, developers and users to obtain below-IP (L1/L2) cellular-specific networking information at runtime from their own smartphones, with no need of operators’ permission to use their network data or using special instruments. It empowers real-time network monitoring, detection, predication and analytics on the device end and thus helps to boost cellular networking performance, efficiency, reliability, and so on.
- Website: http://www.mobileinsight.net
- Github: https://github.com/mobile-insight
- Tutorial (videos and slides): http://www.mobileinsight.net/tutorial-video.html
- First-use doc tutorial: http://www.mobileinsight.net/tutorials.html
MI is suitable to any student, developer and researcher in networking. Participants are required to have
- Basic networking background (TCP/IP network stack, L1/L2, control-/data-plane, throughput, RTT, packet loss)
- Python programming (core and sample codes in Python)
Participants are encouraged to have the following skills or experience
- Linux or Mac OS x (all-in-one VM in Linux) and shell commands
- Basic cellular network knowledge in mobility support (handover)
- Basic Android programming (optional, mainly for fancy UI and widgets)
Our Hackathon challenge is to understand mobility support over LTE. It consists of three progressive projects which can be chosen by every participant based on their interests.
We will use MI to collect RRC-layer messages and extract cellular network mobility support related information including handover behaviors, mobility configurations, measurement reports etc, and MAC-layer information including radio signal measurement for serving and neighboring cells.
We will analyze the information collected in problem 1 and visualize them on desktop. We will monitor measurement reports and mobility configuration changes overtime and link them to the handover behavior. We will display our analyzed results which could potentially lead to handover prediction (optional).
We will try to identify unexpected handover behaviors (e.g., handover failure, ping-pong handover, poor radio connection, etc.) and find their cause (e.g., dynamic radio environment, unwise handover decision, improper mobility configurations, etc.).
Note 1: MI supports two main use modes: (1) Mobile only: run it on phone only (data collection and online analytics), and (2) Mobile+desktop: run mobile app to collect data at runtime but upload data to the desktop for offline data analytics. Hackathon projects will run at mode 2 (mobile + desktop) while they are easily portable to mode 1 (mobile only).
Note 2: MI can be used to understand more than handover, including but not limited to throughput, latency, loss, network dynamics etc. If you are interested in other problems, please feel free to contact us.
Note 3: You are encouraged to run MI before you come to Hackathon. Please read MI’s website (http://www.mobileinsight.net) and go through the tutorial in 1-3 hours. MI’s mobile app runs on rooted Android phones mainly over Qualcomm chipsets. A large variety of phone models are supported. If you don’t have such a phone, please keep me posted and I will bring several MI-ready phones for Hackathon’s use.
Haotian Deng will participate and supervise all the participants who are interested in MobileInsight at the Hackathon event. If anyone has technical questions in advance, please directly contact him (deng164@purdue.edu) and the MI support team (support@mobileinsight.net).