Execution time frame: 2023.03.01 - 2025.02.28.

Coordinator: Dr. Sonkoly Balázs

Members: 3 seniors, 2 postdocs, 1 PhD student, 2 MSc students

Abstract: The aim of this project is to deepen a preexisting cooperation between a Hungarian and an Italian research group. The members of the Hungarian party come from the MTA-BME Network Softwarization Research Group. The Italian party constitutes of members of CNIT (National Inter-University Consortium for Telecommunications) and Scuola Superiore Sant'Anna, Pisa, Italy. During the project, our goal is to enhance the features of our application deployment and operation platform of which foundations were laid previously by the cooperation of the two research groups. We target artificial intelligence (AI) applications with special resource demand and service level requirements. According to our concept, we develop a software framework that supports the operation of latency sensitive applications. Here, end user equipment connects to edge devices (which are located in relative proximity to user equipment) via low-latency radio links. The platform uses edge devices connected via optical networking solutions that execute artificial intelligence-based applications leveraging serverless management. The serverless concept is rooted at public cloud providers and because of its good, on-demand scaling features, it can significantly improve the resource utilization of devices. This feature is in dire need at the network edge where compute resources might be relatively scarce. In the course of this project, we collect telemetry data from optical networking equipment and network interface cards and, based on these, we enhance the platform with fast application component relocation features that can adapt the application layout to sudden changes occurring in the application's input load. As results of this project, we plan to publish two conference papers and based on those a journal followup.

Acknowledgement: This work is supported by the Ministry of Innovation and Technology of Hungary from the National Research, Development and Innovation Fund through project 2021-1.2.4-TÉT-2021-00058 under the 2021-1.2.4-TÉT funding scheme.

THE AMOUNT OF FUNDING:
2.5 MILLION HUF