A data grid is a set of mutually connected computers at geographically different sites, whose aim is to achieve a good availability of the data. By choosing an adequate replication strategy one wants to achieve such an arrangement of the files in the grid that the performance thereof is as good as possible (with regard to the response time and to the bandwidth used). The diploma thesis deals with two scientific papers of the same authors, wherein they suggest improvements of the dynamic replication strategy Fast Spread, yet it has not been possible to repeat their results. Additionally, the great similarity of the papers gives rise to a strong suspicion of autoplagiarism. In the introduction of the diploma thesis, data grids are presented from a theoretical point of view, then it is described how the testing of the strategies suggested in the papers took place and, subsequently, all established defects of the papers including their (too) great similarity are dealt with.
|
