Sidi Mohammed Kaddour

Team STACK

Inria

Research & Development Engineer, Inria STACK

I build and measure distributed systems that span the edge-to-cloud continuum. My work turns on two questions: how do you see what a geo-distributed application is actually doing, and how do you reconfigure it safely once you know? In practice that means application-level observability with OpenTelemetry and Prometheus, coordination-aware deployment with Ansible and Concerto, and experiments large enough to be worth trusting on Grid'5000.

Before Nantes I completed a PhD in networks and distributed systems at the University of Tlemcen, co-supervised with Bournemouth University, on soft sensors for complex systems — and spent five years building and running an ERP in production, which is where my preference for systems that survive contact with reality comes from.

Distributed Systems Observability Edge-to-Cloud Ansible Kubernetes Grid'5000

Explore projects Selected publications Code Open CV

Google Scholar ORCID DBLP GitHub Inria GitLab LinkedIn

Current work

At Inria (STACK team), I am developing CoAnsible, an extension that brings coordination logic from Concerto into complex distributed deployments.

Latest paper

Application level observability for adaptive Edge to Cloud continuum systems, presented at UCC '25 in Nantes — ACM / arXiv / code.

Technical focus

My work combines OpenTelemetry, Prometheus, Kubernetes, and Ansible to build measurable and adaptive systems, with reproducible experiments on Grid'5000.

Motto

“Fail, learn, adapt, repeat.”

If you’d like to talk about research, engineering, or a possible collaboration, the contact links above are the fastest way to reach me.

news

Dec 01, 2025	📄 Application level observability for adaptive Edge to Cloud continuum systems is published in the UCC ‘25 proceedings. The paper combines OpenTelemetry, Prometheus, k3s, and Chaos Mesh into an SLO-aware feedback loop, evaluated on Grid’5000. arXiv · code.
Sep 17, 2025	📰 Exciting News! I am honored to serve as Co-Web Chair for the 18th IEEE/ACM International Conference on Utility and Cloud Computing (UCC 2025) and the 12th IEEE/ACM International Conference on Big Data Computing, Applications, and Technologies (BDCAT 2025), which will be held in Nantes, France. Looking forward to contributing to the success of these conferences!
Jun 16, 2025	📄 Our paper Multi-provider capabilities in EnOSlib was published at DAIS 2025 (DisCoTec, Lille). It extends EnOSlib so a single experiment description can span several testbeds across the edge-to-cloud continuum. Open access PDF.
Dec 01, 2024	Started a new position as R&D Engineer at Inria (STACK team, Nantes), working on CoAnsible — an extension that brings Concerto’s coordination model into Ansible for adaptive distributed deployments.

latest posts

Dec 01, 2024	Started a New Position at Inria
Dec 01, 2023	Feedback Mechanisms for Edge-to-Cloud Applications at IMT Atlantique

selected publications

UCC

Application level observability for adaptive Edge to Cloud continuum systems

Sidi Mohammed Kaddour, Baptiste Jonglez, and Daniel Balouek

In Proceedings of the 18th IEEE/ACM International Conference on Utility and Cloud Computing (UCC ’25), 2025

Abs DOI arXiv Bib HTML PDF Code

This paper introduces an application-level observability framework for adaptive Edge-to-Cloud systems. The approach integrates SLO-aware feedback loops and fine-grained monitoring using OpenTelemetry, Prometheus, and Kubernetes-based infrastructures. Experimental results show improved adaptability, scalability, and resilience under dynamic workloads and fault conditions.

@inproceedings{10.1145/3773274.3774855,
  author = {Kaddour, Sidi Mohammed and Jonglez, Baptiste and Balouek, Daniel},
  title = {Application level observability for adaptive Edge to Cloud continuum systems},
  year = {2025},
  isbn = {9798400722851},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  url = {https://doi.org/10.1145/3773274.3774855},
  doi = {10.1145/3773274.3774855},
  booktitle = {Proceedings of the 18th IEEE/ACM International Conference on Utility and Cloud Computing (UCC '25)},
  articleno = {58},
  numpages = {6},
  keywords = {Edge-to-Cloud Continuum, Application-level Observability, Service-Level Objectives, Adaptive Systems},
  series = {UCC '25}
}

DAIS
Multi-provider capabilities in EnOSlib: driving distributed system experiments on the edge-to-cloud continuum

Baptiste Jonglez, Matthieu Simonin, Jolan Philippe, and Sidi Mohammed Kaddour

In Distributed Applications and Interoperable Systems (DAIS 2025), 25th IFIP WG 6.1 International Conference, 2025

Abs DOI Bib HTML PDF Code

This paper extends EnOSlib with multi-provider capabilities to support experimental research across the edge-to-cloud continuum. The proposed enhancements enable seamless orchestration, reproducibility, and performance evaluation of distributed systems spanning heterogeneous infrastructures, including synchronized reservations across several testbeds.
@inproceedings{jonglez2025multi, title = {Multi-provider capabilities in EnOSlib: driving distributed system experiments on the edge-to-cloud continuum}, author = {Jonglez, Baptiste and Simonin, Matthieu and Philippe, Jolan and Kaddour, Sidi Mohammed}, booktitle = {Distributed Applications and Interoperable Systems (DAIS 2025), 25th IFIP WG 6.1 International Conference}, pages = {25--42}, year = {2025}, organization = {Springer Nature Switzerland}, doi = {10.1007/978-3-031-95728-4_2}, keywords = {EnOSlib, Reproducible Experiments, Edge-to-Cloud Continuum, Grid'5000} }
FlexScience
Proteus: Towards Intent-driven Automated Resource Management Framework for Edge Sensor Nodes

Shashikant Ilager, Daniel Balouek, Sidi Mohammed Kaddour, and Ivona Brandic

In Proceedings of the 14th Workshop on AI and Scientific Computing at Scale using Flexible Computing Infrastructures (FlexScience ’24), 2024

Abs DOI Bib HTML PDF

Proteus presents an intent-driven framework for automated resource management in edge sensor nodes. The framework translates high-level intents into low-level resource configurations, enabling adaptive, efficient, and scalable management of edge infrastructures.
@inproceedings{ilager2024proteus, title = {Proteus: Towards Intent-driven Automated Resource Management Framework for Edge Sensor Nodes}, author = {Ilager, Shashikant and Balouek, Daniel and Kaddour, Sidi Mohammed and Brandic, Ivona}, booktitle = {Proceedings of the 14th Workshop on AI and Scientific Computing at Scale using Flexible Computing Infrastructures (FlexScience '24)}, pages = {1--8}, year = {2024}, doi = {10.1145/3659995.3660037}, keywords = {Intent-driven Management, Edge Sensor Nodes, Resource Management} }
arXiv
Event Detection for Non-intrusive Load Monitoring using Tukey’s Fences

Sidi Mohammed Kaddour, Mohamed Lehsaini, and Abdelhamid Bouchachia

arXiv preprint arXiv:2402.17809, 2024

Abs DOI arXiv Bib HTML PDF

This paper proposes a Tukey’s Fences-based event detection method for non-intrusive load monitoring (NILM). By combining FFT-based analysis with statistical fences, the approach accurately detects appliance switching events in aggregated electrical signals, achieving high detection accuracy compared to existing methods.
@article{kaddour2024event, title = {Event Detection for Non-intrusive Load Monitoring using Tukey's Fences}, author = {Kaddour, Sidi Mohammed and Lehsaini, Mohamed and Bouchachia, Abdelhamid}, journal = {arXiv preprint arXiv:2402.17809}, year = {2024}, doi = {10.48550/arXiv.2402.17809}, keywords = {NILM, Event Detection, Smart Energy, Signal Processing} }
IJSSCI
Electricity consumption data analysis using various outlier detection methods

Sidi Mohammed Kaddour and Mohamed Lehsaini

International Journal of Software Science and Computational Intelligence (IJSSCI), 2021

Abs DOI Bib HTML

This paper investigates abnormal electricity consumption detection using several unsupervised outlier detection techniques. Isolation Forest, One-Class SVM, and K-means clustering are evaluated on real consumption data to identify unusual patterns. The results demonstrate the effectiveness of these methods in detecting energy anomalies and supporting better energy management.
@article{kaddour2021electricity, title = {Electricity consumption data analysis using various outlier detection methods}, author = {Kaddour, Sidi Mohammed and Lehsaini, Mohamed}, journal = {International Journal of Software Science and Computational Intelligence (IJSSCI)}, volume = {13}, number = {3}, pages = {12--27}, year = {2021}, publisher = {IGI Global}, doi = {10.4018/IJSSCI.2021070102}, keywords = {Outlier Detection, Electricity Consumption, Isolation Forest, One-Class SVM} }