Provenance is a key metadata for assessing electronic documents trustworthiness. Most of the applications exchanging and processing documents on the web or in the cloud become provenance aware and provide heterogeneous, decentralized and not interoperable provenance data. A new type of system emerges, called provenance management system (or PMS). These systems offer a unified way to model, collect and query provenance data from various applications. This work presents such a system based on semantic web technologies and focuses on scalability issues. In fact, modern infrastructure such as cloud can produce huge volume of provenance data and scalability becomes a major issue. We describe here an implementation of our PMS based on an NoSQL DBMS coupled with the map-reduce parallel model and present different experimentations illustrating how it scales linearly depending on the size of the processed logs