Today, most of the applications exchanging and processing documents on the web or in clouds become provenance aware and provides heterogeneous, decentralized and not interoperable provenance data. Provenance is becoming a key metadata for assessing electronic documents trustworthiness and should be considered as first class data. Most of provenance management systems are either dedicated to a specific application (workflow, database) or a specific data type. Moreover, in modern infrastructures such as cloud, applications can be deployed and are executed on several provider infrastructure. So, there is a need to track the provenance of applications between different provenance providers. For these reasons, modeling, collecting and querying provenance across heterogeneous distributed sources is considered as a challenging task. In this paper, we introduce a framework based on semantic web models supporting syntactic and semantic heterogeneity of provenance sources and correlations between multiple sources. This framework is implemented as a provenance management system (or PMS). We focus on the design of a mediator based system allowing to federate distributed PMSs and we present optimization issues related to distributed query processing.