Provenance-CSL - A Provenance Client Side Library

Kurzfassung

Data management is a challenge in both scientific and technical environments. Therefore researchers have developed a special interest in this field. Modern approaches (i.e. Subversion, CVS) already offer authoring and versioning in distributed systems. However this might be insufficient in a vast number of scenarios, where not only the data resulting from a process, but also data which describes the process that generated those results is crucial. For example, if a doctor needs to decide how to treat a patient, he must have access to the patient’s data. Moreover he needs to know how the data was obtained (which tests were made), how old it is (when the tests were made), by whom it was provided (which doctor treated the patient in the past) and so on. This means that not only the patient’s condition, i.e. his bloodpressure, needs to be documented, but also the process which led to the data. This process is called the Provenance of the data. Meta-data describing such a process is the process documentation. There are many more scenarios where the process of acquiring the data might be almost as important as the data itself. The task of collecting a processes documentation might be quite easy in simple local systems, but becomes rather difficult in distributed environments. The EU-Project Provenance aims at the development of an open architecture which enables grid-applications to collect and collaborate process documentation in such environments. The general design of the Provenance-architecture is built around a Provenance store. Applications, or actors, which are part of a process may record P-Assertions, which are the atomic units of the process documentation, on the Provenance store. They may query the Provenance store for certain P-Assertions as well. In order to ease the process of developing applications which make use of Provenance, a client side library, which offers a simple API for interaction with Provenance stores, needs to be developed. This report describes the design of a Provenance-architecture, provides details of the API specification and presents the implementation of a Provenance client side library.