Chunyi Peng

My research aims to innovate mobile network access in 4G/5G and beyond to support emerging demands on new functionalities, performance, reliability and security. I am interested in exploiting and developing a multi-disciplinary solution that incorporates machine learning, data science, distributed systems, computing theory, cybersecurity to mobile networking. Technology-wise, my approach integrates three core components in a learning-reasoning-acting cycle: (1) I develop a data-driven approach to learn what happens at runtime; This exposes “closed” cellular network operations to the research community and opens up problems as well as solution space; (2) I apply formal methods to reason why (deepen into its root causes). Combining network verification and empirical validation, I uncover and validate defect flaws, operation slips and configuration errors which have been largely overlooked. I also develop provable and simplified protocols/procedures to meet demanding KPIs. (3) I exploit unexplored device power to act in time and solve problems given hints from the above learning and reasoning procedure. Through turning passive clients to more proactive roles, we empower end device intelligence for better performance, reliability and security.

Currently, I am particular interested in the following problems

Some problems overlap. For example, techniques developed in AIM can boost mobility support and enhance network security.

AIM: Amplify Intelligence in Mobile Networked Systems

 

I envision the future mobile networked system to be featured by its substantially higher degree of intelligence, which offers both high performance (to the customizable services for each mobile user and thing) and built-in resiliency and security against component failures and malicious attacks. This project particularly tackles two challenges in mobile networking research: (1) black-box operations and (2) high complexity (lack of verification) in network infrastructure and operations.

In this project, I seek to enhance intelligence to the signaling subsystem, which is responsible for critical network control utilities such as radio resource control, mobility management, service quality and security. Specifically, our mission is not to expand hard power (like many wireless advances over wider spectrum for higher bandwidth and larger capacity) but to exploit soft control and management power to fulfil network potential and eventually deliver quality services to end devices.

This project is supported by NSF under grants CNS-1750953 (CAREER), CNS-1748630 and CNS-1423576.

Publications

Software and Release

Better Mobility Support in Cellular Networks

 

Mobility support is deemed a fundamental service for the next-generation Internet. The current cellular network is the only large-scale infrastructure that successfully provides wide-area, ubiquitous mobility support in reality. With the explosive growth of smartphone devices and the surge of mobile data traffic, cellular networks have been evolving into an increasingly heterogeneous networked system. As a result, managing mobility becomes challenging yet rewarding.

This umbrella project covers three sub-projects: (T1) single-carrier mobility support (mainly on handoff configuration in the management plane), (T2) multi-carrier access (primarily in Google Project Fi) and (T3) extremely high-speed mobility support. In T1, we seek to study the configuration issues on mobility management of cellular networks, in order to ensure desirable mobility support. We start with two structural properties: stability and reachability. Stability implies no persistent oscillation loops during constant network conditions, while reachability denotes no access black hole (e.g., certain cells or even a given mobile technology (e.g., 4G) cannot be reached by the device). We then extend our study to its performance impacts and devise solutions to enhance mobility support with adaptive configurations. In T2, we extend our single-carrier study to multi-carrier access, with a focus on inter-carrier switching. We reveal unanticipated problems in policy conflicts and improper configurations and devise solutions for seamless and quality mobility support that seemed within one carrier. In T3, we investigate the (negative) impact of extremely high speed (much frequent handoffs), which arises with advanced transport facility and 5G requirements. We quantify poor performance under extremely high speed and their root causes and shed design insights for new mobility solutions tailored to extremely frequent handoffs.

T1 is partly supported by NSF under grants CNS-1749049 and CNS-1526985 (more).

Software and Release

Publications

Evoling Mobile Network Security

 

Changes are unchanging sources of insecurity in an evolving system. Along with new features or revised functions for emerging demands, cellular networks unfortunately become vulnerable without prudent design and operations. These loopholes either lie in open attack interfaces which were closed but now exposed to attackers, or no bullet-proof shield among multiple components, procedures and functions. In this project, we focus on the new breaches associated with newly emerging functions and/or substantial updates like 4G voice/SMS support, IoT, data billing. Through these showcases, we identify loopholes in operational carriers (primarily top-tier US carriers), devise proof-of-concept attacks (validated and assessed in controlled and responsible experiments) and propose defense schemes and immediate fixes.

Securing mobile data charging is supported by NSF under grants CNS-1753500 and CNS-1422835 (more).

Voice/SMS (or IMS) security is supported by NSF under grants CNS-1749045 and CNS-1528122 (more).

Industry impacts: Our results on mobile data billing and voice call security have been immediately adopted by US carriers (AT&T, T-Mobile, Verizon). Most threatening attacks have been fixed with nationwide network upgrades.

CEIVE: Callee-Only Inference and Verification Against Caller ID Spoofing

Publications

Software and Release

Novel Mobile Communication, Sensing and Applications

We exploit new wireless media (primarily light and sound) to enrich better, easier, and more convenient in-proximity communication, interactions and sensing for mobile, such as (imperceptible) screen-to-camera communication, indoor positioning/sensing over light or sound. We leverage the gap between human and machines in perception and learning to address user needs in specific use scenarios.

Publications

Past Projects

Acknowledgement

I gratefully acknowledge ongoing and past research support from NSF and Adobe.