Newsletter #1 (7/2023)
The Project at a glance
MLSysOps is a HORIZON EUROPE funded project with a consortium of 12 partners from 8 countries and it has a duration of 36 months.
MLSysOps stands for Machine Learning for Autonomic System Operation in the Heterogeneous Edge-Cloud Continuum.
The project aims to design, implement and evaluate a complete AI-controlled framework for autonomic end-to-end system management across the device-cloud-edge continuum. MLSysOps employs a hierarchical agent-based AI architecture to interface with the underlying resource management and application deployment/orchestration mechanisms of the continuum.
Energy efficiency and utilization of green energy, performance, low latency, efficient, and trusted tier-less storage, cross-layer orchestration including resource-constrained devices, resilience to imperfections of physical networks, trust, and security, are key elements of MLSysOps addressed using ML models.
MLSysOps will be evaluated through two well-defined use cases utilizing cloud, smart, and deep-edge infrastructures:
- Precision Agriculture
- Smart Cities
A few words from our Coordinator
It is with great pleasure that I welcome you to the first edition of the MLSysOps project newsletter. As the Coordinator of this ambitious European project, I am honored to share our collective journey toward advancing the field of Autonomic System Operation in the Heterogeneous Edge-Cloud Continuum using Machine Learning (ML) methodologies.
At the core of MLSysOps is the design, implementation, and evaluation of a framework for autonomic end-to-end system management across the entire cloud-edge-IoT continuum. This framework will leverage a hierarchical agent-based ML architecture, serving as an interface to the underlying resource management and application deployment/orchestration mechanisms of the continuum. To achieve adaptivity, MLSysOps will incorporate continual ML model learning in tandem with intelligent retraining during application execution. The project emphasizes openness and extensibility, employing explainable ML methods and an API for pluggable ML models.
MLSysOps addresses essential elements such as energy efficiency (incl. the use of green energy), performance, latency, efficient and resilient storage, resource-constrained devices and network connectivity, trust, and security, utilizing ML models as solutions.
The framework architecture of MLSysOps dissociates management from control and seamlessly interfaces with popular control frameworks across different layers of the continuum. To validate and refine the framework, the project will employ research testbeds as well as two real-world application-specific testbeds in the domains of smart cities and smart agriculture. These testbeds will also provide the necessary system-level data for training and validating the ML models. Additionally, realistic system simulators will enable scale-out experiments.
The MLSysOps consortium comprises a balanced blend of academic/research institutions and industry/SME partners. This collaborative effort brings together the requisite scientific and technological expertise, ensuring successful implementation and impactful outcomes. By combining cutting-edge research, innovative methodologies, and collaborative partnerships, MLSysOps aims to pave the way for more efficient, resilient, and intelligent systems in domains critical to our society’s progress.
This newsletter, which will be published every 6 months, provides a glimpse of the progress and achievements of MLSysOps. We invite you to join us on this exciting journey as we shape the future of adaptive machine-learning systems. By all means, feel absolutely free to contact us for more information about the work done on the project.
University of Thessaly
Our project was launched in January 2023. The Consortium traveled to Volos, Greece, to attend the kick-off meeting hosted by the University of Thessaly.
Over the course of two days, the partners engaged in a comprehensive and constructive dialog concerning the project’s goals and operational strategy. This included in-depth talks about work packages, the technologies and use cases to be implemented and the necessary actions for securing the project’s success, and established the basis for a highly promising and productive partnership.
The 1st International Workshop on Machine Learning for Autonomic System Operations in the Device-Edge-Cloud Continuum (MLSysOps 2023) will be held in conjunction with the International Conference on Embedded Wireless Systems and Networks (EWSN 2023) on September 25, 2023 in Rende, Italy.
The goal of the workshop is to bring together a community of researchers and practitioners who study problems at the intersection of AI/ML, autonomic and cognitive computing, Device-Edge-Cloud continuum, distributed system operation, and resilient application deployment.
The topics of interest include the autonomic management and control of D-E-C continuum, ML/AI-driven approaches for system operation of dynamic, large-scale, heterogeneous continuum systems, agent-oriented architectures for D-E-C continuum orchestration, cognitive computing models in the edge-cloud continuum, AI/ML-based management of computing, networking and storage resources in the continuum, green, resource-efficient techniques for system operation, AI/ML-based trust and security methods in the edge-cloud continuum, network and system simulators for the D-E-C continuum, AI/ML-based application deployment in continuum systems.
Participation to the Consultation & Concertation event in Brussels
MLSysOps actively participated to the “EU Concertation and Consultation on Computing Continuum: From Cloud to Edge to IoT” event organized by EUCloudEdgeIoT on May 10-11 2023 in Brussels. Thanks to the great gathering of leading professionals, we put the basis for interesting collaborations with other projects of the Cognitive Cloud cluster.
ACACES23 Summer school
MLSysOps participated in ACACES23 Summer School, organized by HiPEAC on July 9-15, 2023. A poster presenting a part of our work, related to node-level ML-based autonomic operation, gave us the opportunity to get valuable feedback from the research community and expand the project’s network.
- V. Barbuto, C. Savaglio, M. Chen, G. Fortino, “Disclosing Edge Intelligence: A Systematic Meta-Survey”. Big Data and Cognitive Computing, 7(1):44, 2023
- G. Polychronis, S. Lalis, “Flexible Computation Offloading at the Edge for Autonomous Drones with Uncertain Flight Times”, 19th IEEE International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT), accepted (to be presented/published).
- M. Loaiza, C. Savaglio, R. Gravina, D. Chatzopoulos, S. Lalis, “Agents in the computing continuum: the MLSysOps perspective”, The 1st International Workshop on Machine Learning for Autonomic System Operation in the Device-Edge-Cloud Continuum (MLSysOps 2023), accepted (to be presented/published).
- A. Patras, F. Pournaropoulos, N. Bellas, C. D. Antonopoulos, S. Lalis, M. Goutha, A. Nanos, “A Minimal Testbed for Experimenting with Flexible Resource and Application Management in Heterogeneous Edge-Cloud Systems”, The 1st International Workshop on Machine Learning for Autonomic System Operation in the Device-Edge-Cloud Continuum (MLSysOps 2023), accepted (to be presented/published).
- T. Aslanidis, A. Chouliaras, D. Chatzopoulos, “Reinforcement Learning Techniques for Optimizing System Configuration on the Cloud: A Taxonomy and Open Problems”, The 1st International Workshop on Machine Learning for Autonomic System Operation in the Device-Edge-Cloud Continuum (MLSysOps 2023), accepted (to be presented/published).
- F. Pournaropoulos, C. D. Antonopoulos, S. Lalis, “Supporting the Adaptive Deployment of Modular Applications in Cloud-Edge-Mobile Systems”, The International Conference on Embedded Wireless Systems and Networks (EWSN 2023), accepted (to be presented/published).
- T. Aslanidis, M. Koutsoubelias, S. Lalis, “Transparent Handover of Automated Drone Missions between Edge-based Control Stations”, The International Conference on Embedded Wireless Systems and Networks (EWSN 2023), accepted (to be presented/published).
The “MLSysOps” Project is funded by the European Community’s Horizon Europe Programme under Grant Agreement #101092912.
Copyright © 2023, MLSysOps EU Project, All rights reserved.