BrIAS Seminars: Giang Nguyen

13.03.2025

BrIAS Fellow Prof. Giang Nguyen

Softwarized Networks for Networked and Sustainable Cyber-Physical Systems

Abstract: Future use cases, like robotics, healthcare, and autonomous vehicles, need cyber-physical systems (CPSs). Such a system integrates physical and computing components to create a perceived real-time interaction between the physical and digital worlds. Central to CPSs are computer networks connecting heterogeneous technologies such as sensors and actuators to sense, analyze, and respond to physical phenomena in real-time. Such networks require flexibility and resilience to adapt to changing conditions and unexpected events, such as equipment failures.

Conventional networks with specialized networking devices are costly and inflexible for new concepts like mobile edge clouds (MEC). Softwarized networks leverage virtualization technologies to develop and deploy virtualized network functions (vNF), which are software running on commodity hardware. This shortens innovation cycles and allows for elastically adjusting computed resources. The challenge is minimizing latency due to packet processing in software.

In this talk, I will present three research highlights. The first study explores the feasibility of vNF for real-time human-machine co-working. We develop, implement, and evaluate CALVIN, a low-latency management framework for vNF, and assess its performance. The second study transforms the centralized ICA (Independent Component Analysis) algorithm into a decentralized version to leverage the softwarized networks, accelerating the Blind Source Separation of audio signals, reducing transferred data, and thus saving energy. The third study concerns StateOS, a software framework to manage network function processing results (or states). The conventional approach is to replicate the whole application memory, which is inefficient and time-consuming. StateOS applies different strategies for storing and transferring connection and action states, reducing the volume and duration of state transfer. Last but not least, I will briefly summarize the latest outcomes from the collaboration at BrIAS.