I have been enrroled at the Institute of Mathematics and Statistics (IME) of the University of São Paulo (USP) in the undergraduate degree of Computer Science. As my conclusion work, I developed the following research project in the field of Cloud/ Edge Computing, Scheduling of jobs and Smart Cities:
Edge Computing Platforms with Smart Heaters: Investigating Job Allocation Policies, Describing and Modeling Temperatures
Under the supervision of:
- Prof. Denis TRYSTRAM,
- Prof. Alfredo GOLDMAN,
- Dr. Danilo CARASTAN
This project was built in two parts, under the following two main topics:
-
Smart Heaters as an Edge Computing Platforms: Investigating Job Allocation Policies
The first part was built during an internship where I was enrroled in the second year of the Master's Program of the University of Grenoble Alpes (UGA). For that I was sponsered by a scholarship of innovation and entrepreneurship from AUSPIN (Agência USP de Inovação). This project was supervised by the professor Alfredo Goldman, from the USP sides, and the professor Denis Trystram, from the UGA sides.
This project was focused in the development of a simulator for Edge platforms, focused in the job allocation policies, and it was based on a real use cases, the Qarnot Computing, which is a company that offers Smart Heaters, that by one side heat general users environments, by other side produced the heat needed by running HPC Jobs for Cloud Computing users. For that, I perfomed the following topics:
- I worked on the development of an Edge Platform on top of the simulator, based on an use case.
- I implemented and compared five scheduling algorithms.
- I developed an experimental methodology to improve the use and understand the platform.
At the end of the program, I defended my master thesis based on this research project and became a Masters' of Science (MSc.) by the University of Grenoble Alpes on February 2020. Please, find the documents produced so far:
Edge Computing Platforms with Smart Heaters: Describing and Modeling Temperatures
In order to finish my undergraduate studies at the University of São Paulo, the second part of this research is a continuation of the first part developed at the University Grenoble Alpes, as my Masther Thesis. This part was also under the suppervision of professor Alfredo Goldman and professor Denis Trystram, with the addition of the Dr. Danilo Carastan.
What was at first focused in the development and analysis of an simulated Edgde Platform based on a real use case, now is focused in the data analysis from the real logs of its use case. Regarding the novelty presented by the company Qarnot Computing, this part of the work was focused in the analyses from their real logs, in order to go one step further to build a temperature model to be added in the simulator to then improve its temperatures predictions. For that I did:
- An analysis descript of the Qarnot logs
- A categorization algorthms to group the Qarnot's Smart Heaters.
Please, find more details in the full document below. Notice that you will follow the manuscript from the first part, merge into this next one:
- Undergradute Thesis Manuscript [download here]
- Undergradute Thesis Repository: https://gitlab.com/andersonandrei/usp-tcc