For the analysis of geospatial data, statistical spatio-temporal aggregation (SSTA) is a frequently used functionality. Possible applications include information extraction, data fusion, generalization and schema transformation. On the basis of spatial and temporal references, SSTA transfers thematic attribute values into a coarser spatio-temporal resolution using descriptive statistical operations. Reusable geoprocessing functions are increasingly offered in Spatial Data Infrastructures (SDIs) through standardized interfaces using mainstream network technologies. Due to the variety of use cases, a Web service for SSTA promises a high potential for reuse. However, offering functionalities for geoprocessing within an SDI raises some challenges. The aggregation process itself, as well as the encapsulation of the aggregation functionalities within an SDI, require thorough design considerations. This article addresses these challenges by developing a generic framework for SSTA services within an SDI. A prototype of the framework is realized using several open-source software components. Following a modular approach, the communication between those loosely-coupled components is enabled through open, standardized interfaces. The result of the proposed work is a framework that enables users to easily perform a statistical ad-hoc aggregation on distributed spatio-temporal data.