Data Warehousing for INventory management DWIN is a production project at Telcordia aimed at providing telecommunications service providers with decision support functions for inventory control and monitoring. In this paper, we report some interesting issues related to the design of the data warehouse. Speci cally, we will discuss the issues of slow-evolving fact, transaction-oriented fact table, and large dimensions. We also propose the concept of virtual data cubes and show its usefulness. We address these issues through a data mart case study and present benchmarking results. Finally, based on the experiences learned, we discuss potential research issues that may bene t the data warehousing and OLAP practice.