Subsurface modeling workflows are complex, data intensive, and iterative, requiring many different software packages to repeatedly exchange data. Improving the speed and accuracy of data exchange in model development can improve the speed, accuracy, and reliability of analyzing these models, ultimately supporting better decisions for asset development and economics over the life of a field. RESQML is an XML/HDF5-based, data-exchange specification that allows data to be transferred efficiently between the many different software packages used in subsurface modeling workflows. Key Version 2 enhancements include: a richer, more detailed RESQML data model with more domain data-objects (e.g., wells), and the ability to group related data-objects (e.g., faults, horizons, grids, etc.) and exchange them as a complete model. While these enhancements support more accurate and reliable models, they also presented challenges that were solved using new technological approaches new to RESQML. To design the richer yet more complex data model, the RESQML Special Interest Group (SIG) has moved from a hierarchical view of the data model to an entity-relationship view. The SIG is now using UML modeling tools to visualize the data model and produce both class and instance diagrams. The class diagrams are then used to generate XML schema definitions in a more automated and consistent manner. The Open Packaging Conventions_a container-file technology that stores a combination of files and their relationships to form a single entity for transfer in one compressed (ZIP) document format_are being used to group together independent data-objects (files) as a complete model. This paper presents an overview of RESQML Version 2 enhancements and capabilities, and explains how these new technology solutions are being used and their impact on subsurface modeling, the design process, future maintenance, and in helping software developers integrate RESQML into their own products.