Eddy's website

Need to know something about me?

Latest news

October, 2023
Cléa. 10km Run'in Lyon. 00:58:11

June, 2022
Teo. 100 NL (50m) 01:02:42

December, 2021
Teo. 50 NL (25m) 00:28:97

December, 2021
Cléa. Sauts Long. 3m98

November, 2021
Cléa. 50m plat. 7"92

October, 2021
Teo. 10km Run'in Lyon. 00:46:34

Links:

Location: / / Teaching / Cours M2 CR17 (2022-2023) Language: en
 

Large scale sustainable distributed resource management


The Team

The Teaching Team:

  • Eddy Caron (ENS-Lyon)
  • Laurent Lefevre (Inria)



General Presentation

In this course given by three teachers we will discover the IT resource management through different point of view. Some ideas are given in the following of this page.

Lecture Schedule

Part 1: Infrastructure

In this Part we will discover different kinds of distributed infrastructure (PaaS, IaaS, SaaS, VNF, etc.)

  • Clouds platform
  • Edge / Fog platform

Part 2: Middleware and ressource management

This part introduces three main topics.

  • Middleware for Cloud environment
  • Data Management for Cloud and Edge Infrastructure
  • Deployment Technics (with practical session)

Part 3: Sustainability in resource management for large scale distributed systems - Eco design and energy leverages on scalable systems

After a survey of why energy is one of the main limitating factor for the design of large distributed connected systems (LDCS) (Clouds, Fog, Datacenters; HPC systems), we will introduce the concepts of energy management and energy efficiency in LDCS. We will focus on various LDCS energy levers, models and algorithms allowing energy reduction at large scale. We encompass energy aspects with full impacts analysis of large ICT systems. We continue by presenting on-going research done in this domain within ENS-Lyon. We complete this part with a Practical sesssion on Silecs platforms dealing with energy profiling and analysis, eco-design and energy efficiency metrics.

Part 4: Dynamic Systems

Abstract: Traditional distributed computing relies on a number of implicit assumptions that are not always realistic in modern networks. Notably, distributed systems are usually assumed to be static, meaning that system membership and the communication graph do not change over time. As a result, many existing distributed system models are not applicable to wireless networks. New, more complex models have been designed for dynamic distributed systems, and existing algorithms have been adjusted or replaced to function in these new environments. In this part, we will study these new models and algorithms.

  • Distributed algorithms for dynamic systems

Evaluation

Evaluation will be based on practical session, article reading and writing a report.