Semantic Business Process Management (SBPM) utilizes semantic technologies to achieve more automation throughout the BPM lifecycle. An integral part of the SBPM infrastructure is a Semantic Business Process Repository (SBPR), which is used for storage and management of business process modeling artifacts. As in SBPM business process models are based on process ontologies, the SBPR has additional requirements towards support of reasoning and querying capabilities. In this paper, we first identify the functionalities the SBPR has to provide. We then evaluate different approaches on how process models can be stored and queried efficiently by taking the semantic information into account. Finally, we present the overall architecture for the SBPR.