This talk describes a new system Quilt which supports security relevant querying across heterogeneous network data sources. Using a two-level architecture Quilt is able to provide for integrated temporal matching across varying data without either centralizing the data or forcing large-scale conversions away from the native data formats. This allows for scalable deployments of Quilt and for extremely expressive queries exploiting the native query systems for each form of data. Quilt will also support abductive matching which permits it to deal with incomplete data capture in a consistent and controlled manner. The presentation will detail Quilt's architecture and query capabilities and demonstrate a prototype of this system.