Towards a Privacy-Aware Communication for Disaggregated Fog Platform

Fog computing is an intermediate layer between edge devices and cloud computing that provides execution support to minimize application delays. Fog aims at improving latency and the quality of experience. Unlike legacy monolithic systems, in which resources are tightly coupled, disaggregated architectures aim to improve resource utilization, scalability, and energy efficiency, while supporting various large-scale workloads. This paper proposes a data communication fabric for disaggregated hardware-architecture-based fog platforms. Particularly, it integrates an often-overlooked component, i.e., privacy as a core design principle. The proposed architecture focuses on distributed and decentralized network traffic management using a vector packet processing platform. To improve latency and privacy, we have presented a target disaggregated hardware and a communication mechanism for efficient data packet management at local and global levels. We demonstrated that vector-packet-processing-based switches outperform traditional Linux bridging, achieving higher throughput (up to 1.7 Gbit/s vs. 0.9 Gbit/s) and lower latency (under 1 ms kernel-space round-trip time).