| | 9 | In this quarter we brought together the multiple Adaptors that have been developed for different types of GENI experiments within the easy-to-use web interface of the NetKarma Portal and presented a tutorial to GENI experimenters that allowed them to see how they can capture provenance describing their experiments and how visualizing the provenance of their experiments using the NetKarma visualization tools can assist in both communicating the results of their experiments and assist them in analyzing their results. |
| | 10 | |
| | 11 | A basic tenet of the Karma provenance system and NetKarma in particular has been the capture of provenance while minimizing the effort required of scientists and researchers to instrument their code or manually enter metadata. From our experience with capturing provenance and metadata, if the process cannot be automated, the capture will not be done since it is an extra burden on researchers. The NetKarma project developed an "Adaptor" approach that scavenges provenance and metadata annotations with a goal of minimizing the effort of GENI experimenters. Early efforts focused first on experiments run using GUSH, and with minor changes that Jeannie Albrecht's group at Williams College incorporated into GUSH, we were able to extract provenance from GUSH logs and generate provenance graphs through NetKarma using the Open Provenance Model (OPM). To enable the visualization of the provenance graphs being generated, we extended the widely used third-party Cytoscape visualization tool with NetKarma plug-ins for retrieving provenance graphs from a NetKarma server and applying layout and formatting capabilities that convert the OPM graphs retrieved into visualizations experimenters can use. |
| | 12 | |
| | 13 | This approach was extended at GEC__ to also extract additional annotations about an experiment from the GMOC database, and following GEC__ we made a NetKarma server available to GENI experimenters to allow them to upload log files to the remote NetKarma server at Indiana University and visualize the resulting graphs in Cytoscape. To demonstrate the applicability of the NetKarma approach to a wider gamut of GENI experimenters, we worked with the GPO to identify experiments that could benefit from Provenance, and starting with NS2 experiments performed by researchers at Clemson that were presented in a demo and poster at GEC13, we have now extended that effort to include ORBIT experiments and provenance generated based on manifest files such as the XSP experiment used in the NetKarma tutorial at GEC14. |
| | 14 | |
| | 15 | In GEC14 we brought all of the components of NetKarma together in an easy-to-use NetKarma Portal that enables experimenters to define a collection for an experiment and add log files and other data to the Portal using a drag-and-drop interface. On the back-end, the Portal uses the NetKarma Adaptors and a NetKarma server at Indiana University to extract provenance and metadata; generating OPM graphs that can be visualized either through the portal using the web-based version of Cytoscape or downloaded and visualized using the Cytoscape desktop version with the NetKarma plug-ins. |
| | 16 | |
| | 17 | The ORBIT experiments shown at GEC14 illustrate the initial goals of the NetKarma project in that experiments using the ORBIT Traffic Generator (OTG) can capture detailed provenance, download the graph from the Portal, and visualize it with no additional instrumentation. At GEC 14, ____ |
| | 18 | |
