5) Provenance information is collected so that the user can know which executable has previously been used with what kind of data and what were the results.
6) Results can be visualized and/or forwarded to a next step in the workflow (e.g. pre-processing, processing).
4) Execution - jobs are executed in parallel on clusters or individual servers anywhere on the grid. This is a very complex mechanism since executables have to be matched with available free computers with suitable operating systems, the executables, libraries and data have to be copied in all of the designated locations, and after execution, results are collected and clean-up operation is performed.
5) Provenance information is collected so that the user can know which executable has previously been used with what kind of data and what were the results.
6) Results can be visualized and/or forwarded to a next step in the workflow (e.g. pre-processing, processing).
Feel free to comment.
vstankovski 2 years ago
4) Execution - jobs are executed in parallel on clusters or individual servers anywhere on the grid. This is a very complex mechanism since executables have to be matched with available free computers with suitable operating systems, the executables, libraries and data have to be copied in all of the designated locations, and after execution, results are collected and clean-up operation is performed.
vstankovski 2 years ago
Steps of the workflow:
1) Selection of available executables from a grid-wide repository (i.e. the executables can be found anywhere in the distributed grid environment)
2) Selection of data for the data mining process. Again, the unit alows to connect data source from anywhere on the grid.
3) Setting up parameters of the grid application (see technical paper for details).
vstankovski 2 years ago
This application is executed on a grid (clusters, individual servers running windows and linux OS) spanning few EU countries.
vstankovski 2 years ago