In this presentation, we describe our two-year journey at Princeton University Library, developing and implementing a new ecosystem of support for research data, known as Princeton Data Commons (PDC). Our previous solution was a heavily customized legacy DSpace instance that housed a variety of data and was no longer sustainable to maintain or feasible to extend. For Princeton Data Commons we decided to use a different model and to build a solution that was specific for research data and the ever- evolving needs and expectations of researchers. Princeton Data Commons is in fact a set of applications dedicated to specific workflows in research data stewardship. We have implemented PDC Describe to allow researchers and curators to deposit new datasets, ensuring they are properly described using metadata standards appropriate for research data. Data ingested into PDC Describe flows into PDC Discovery, another new application we developed for public search, discovery, and download of datasets curated through PDC Describe. In this presentation we’ll talk about the the issues that arose over time in our legacy solution, the features that we added in our new services, how we rolled it out, our uniquely hands-on enhancement-focused approach to data migration from the legacy system to the PDC ecosystem, how we are dealing with larger (and larger!) datasets, how we decided what to build versus what to outsource, some of the challenges that we encountered as the landscape for research data continued to evolve, how we worked with our stakeholders through the entire process, and our plans for the future.