Data modeling for the virtual observatory

The data modeling effort has played a key role in the Virtual Observatory project, and contributed to the effort to build a common reference framework to describe the necessary information attached to astronomical data: the metadata. Such metadata describe the observing parameters and characterize and qualify the observed measurements. These pieces of information are produced and stored in project archives. Standardizing a homogeneous representation of metadata allows uniform discovery and use of the data in the Virtual Observatory infrastructure. This paper describes the context of data modeling in the VO architecture and shows how data models support requirements on the data access layer and for applications development. How the modeling process has been undertaken is explained with a short overview of the different data models. We also discuss in some detail the lessons learned in this modeling and standardization effort.