PACD: A Bitmap-Based Framework for Processing XML Data

Current XML/RDBMS storage models and query processing technologies are reviewed in this paper, leading to the identification of query expressiveness and performance limitations. A novel serialized XML query processing framework is proposed to address these. The proposed query processor (called PACD) is based on a bitmap representation for XML's structural relationships. XPath axes, plus their extension (i.e. "Next" axis) for accessing the document order, are translated to sparse matrices allowing data compression, query complexity reduction and XML updates relaxation. Experimental results, outlined in this paper, show promising performance improvements over conventional techniques in a wide range of query types.