EC-Co-Flows: Enabling Dynamic Edge-Cloud Collaborative Computing Flows

• Applicant: Professorin Dr. Fang-Jing Wu, Ph.D.
• Subject Area: Security and Dependability, Operating-, Communication- and Distributed Systems ; Communication Technology and Networks, High-Frequency Technology and Photonic Systems, Signal Processing and Machine Learning for Information Technology
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 511568981

The evolution of communication and computing techniques tremendously takes place as the softwarization and virtualization techniques are rapidly developed. The emerging paradigms allow a job requested by user applications to be executed in software-based service functions (SFs) instead of hardware. Therefore, this project considers a generic architecture by taking both computing and communication resource allocation issues into account in order to host dynamic jobs from diverse applications. A job consists of a sequence of vertical and horizontal tasks to access an application. These vertical tasks are executed by network SFs in the protocol stack for communication purposes (e.g., a layer-3 gateway followed by a layer-4 traffic monitoring). The horizontal tasks in the sequence are executed by data analytical SFs at the application layer for computing purposes (e.g., an object detection task followed by an annotating task in an AR application). Each task can be executed on any computation resource in the network (i.e., either an Edge or a Cloud server) by dynamically deploying required SFs. This project will address resource allocation issues from three perspectives: SF placement, computing flow scheduling, and SF replacement, to reduce computing and communication costs. The generic architecture proposed in this project, called "EC-Co-Flows", will enable collaborative computing flows between Edges and Clouds. To achieve this goal, this project proposes 5 work packages (WPs): WP1 to design and model EC-Co-Flows architecture, WP2 to schedule horizontal SF flows, WP3 to enable vertical SF Flows based on reinforcement learning (RL), WP4 to design SF replacement algorithms, and WP5 for integration and evaluation of EC-Co-Flows. The project will make technical breakthroughs by enabling the following capabilities in the proposed EC-Co-Flows: (1) efficiently allocating communication-computing resources, (2) adaptively hosting dynamic computing flows, and (3) enhancing resilience with elastic communication-computing architecture. We expect that the outcomes of this project will flourish scientific activities and the technical progress in next-generation communication and computing techniques.

• DFG Programme: Research Grants