SORS: Management in the Cloud-Edge Continuum

Severo Ochoa Research Seminars  WICS  Women BSC  BSC Events  Computer Sciences Seminars 

Date: 05/Jun/2025 Time: 14:30

Place:

[HYBRID] Room 1-2-1, BSC main building and Online via Zoom

Primary tabs

AI4S registration link

  • Title: Towards Efficient Network and Service Management across the Cloud Continuum in the Age of AI
  • Speaker: José Santos. Postdoctoral Researcher in the IDLab Research Group at Ghent University - imec, Belgium
  • Abstract: The rapid adoption of Artificial Intelligence (AI) is reshaping the design and operation of cloud and edge infrastructures. In this evolving landscape, containerization has become essential for managing distributed microservice-based applications, enabling greater flexibility and efficiency. Simultaneously, the emergence of the Cloud Continuum—from centralized data centers to highly distributed edge environments—introduces new challenges for the seamless orchestration of both computational and networking resources. This talk will explore the pressing challenges and opportunities in achieving efficient network and service management across the cloud continuum in the AI era. It will highlight recent advances in container orchestration and explore how AI-driven techniques, particularly Reinforcement Learning (RL), can enable proactive and adaptive orchestration strategies. In addition, this session will emphasize how AI workloads, such as inference and distributed training, introduce diverse network demands—ranging from low-latency requirements to high-throughput communication patterns—that can strain existing cloud infrastructures. It will discuss how orchestration strategies and topology-aware deployment can drastically affect network congestion, throughput, and completion time, especially in modern data center networks. The talk will conclude by identifying key open research questions and emerging directions for building sustainable, scalable, and intelligent next-generation cloud and edge systems in the age of AI.
  • Short bio:  José Santos obtained his M.Sc. degree in Electrical and Computer Engineering in July 2015 from the University of Porto, Portugal. He completed his doctoral studies at Ghent University in April 2022, and he is currently a Postdoctoral Researcher in the IDLab Research Group at Ghent University - imec, Belgium. His research interests include Cloud and Fog Computing, IoT, Service Function Chaining, and Reinforcement Learning. He has authored over 40 scientific publications and received several awards, including the imec Ph.D. Excellence Award (2022), the Best Dissertation Award at NOMS 2023, and the FWO IBM Innovation Award (2023) for his doctoral research. 
  • Title: Towards Autonomous Management in the Cloud-Edge Continuum
  • Abstract: Artificial intelligence (AI) applications are becoming pervasive in all businesses. Many efforts have been put by the AI community to enhance the ML algorithms or enable model optimizations to improve the execution efficiency and robustness of applications. On the other side, hardware manufacturers have been continuously working on next-generation equipment to provide more computing resources to support AI, GPU, FPGA, etc. We agree both directions are essential to benefit AI applications, however, for the massive employment of AI applications and heterogeneous hardware in the emerging hybrid Cloud-Edge Continuum (CEC), efforts should be put more in the management layer to explore new solutions to add additional benefits of making use of the resources and improving AI efficiency, and energy efficiency. The presentation will present an intelligent management layer for scheduling and management of AI applications on CEC, which performs the holistic adaptation at different layers. Additionally, we will highlight key architectural components of the autonomous management framework, a practical AIOps platform and its core capabilities like anomaly detection and predictive maintenance, and real-world use cases in the CloudSkin EU project and lessons learned.  
  • Short bio: Peini Liu is a Recognised Researcher at Barcelona Supercomputing Center. She received her M.S. degree in Software Engineering at National University of Defense Technology (NUDT), China, in 2018, and her PhD degree with honer in Computer Architecture at Universitat Politècnica de Catalunya (UPC) and Barcelona Supercomputing Center (BSC), in 2023. She obtained FI grants for her PhD and WISC primer scholarship from IEEE Service in 2022. She has published several articles in high-ranked conferences and journals, and has experience from several national and international (EU) research projects and is currently leading a group in the CloudSkin EU project. Her research interests include virtualization/containerization technologies, cloud-native, resource management and self-adaptive systems. 

Speakers

Speaker:
  • José Santos. Postdoctoral Researcher in the IDLab Research Group at Ghent University - imec, Belgium
  • Peini Liu. Postdoc - Recognised Researcher at Barcelona Supercomputing Center, BSC
Host: 
  • Jordi Guitart. Associate Researcher. Computer Sciences - Data Centric Computing, BSC
  • Ramon Nou. Estalished Researcher. Computer Sciences - Data Centric Computing, BSC