Skip to Main content Skip to Navigation
Theses

Gestion de donnée complexes pour la modélisation de niche écologique

Ndiouma Bame 1
1 BD - Bases de Données
LIP6 - Laboratoire d'Informatique de Paris 6
Abstract : This thesis concerns large scale biodiversity data management. Its objectives are to optimize queries for researchers who have free access to biodiversity worldwide data. These data which are shared by worldwide research laboratories are federated in GBIF data warehouse. GBIF makes accessible its data to researchers, policy makers and general public. With a significant amount of data and a rapid growth of data and users that express new needs, the GBIF portal is facing a double problem of expressiveness of queries and of efficiency. Thus, we propose a decentralized solution for biodiversity data interrogation. Our solution combines the resources of several of remote and limited machines to provide the needed computing and storage power to ensure system responsiveness for users. It also provides high-level query interface which is more expressive for users. Then, we propose a dynamic data distribution on demand approach. This approach which is based on data properties and characteristics of users analysis queries adapts dynamically machines capacities to users demands. Then, we propose a queries optimization approach that adapts dynamically data placement and machines loads according to performances in order to process users queries within deadlines. We experimentally validated our solution with real GBIF data concerning 100 million observation data.
Document type :
Theses
Complete list of metadata

Cited literature [55 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01191682
Contributor : Abes Star :  Contact
Submitted on : Wednesday, September 2, 2015 - 12:17:05 PM
Last modification on : Thursday, April 15, 2021 - 3:06:54 AM
Long-term archiving on: : Wednesday, April 26, 2017 - 2:44:39 PM

File

2015PA066125.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01191682, version 1

Citation

Ndiouma Bame. Gestion de donnée complexes pour la modélisation de niche écologique. Base de données [cs.DB]. Université Pierre et Marie Curie - Paris VI; Université Cheikh Anta Diop (Dakar), 2015. Français. ⟨NNT : 2015PA066125⟩. ⟨tel-01191682⟩

Share

Metrics

Record views

813

Files downloads

1194