Document fragmentation for XML streams based on Hole-Filler model

A document fragmentation policy was presented by taking advantage of document object model (DOM) for XML, and a corresponding fragmentation algorithm was designed according to the element fan-outs, to solve the problem of document fragmentation for XML streams based on Hole-Filler model. A tag-based document fragmentation algorithm built on DOM-based algorithm was then proposed to determine document filler points by fragmenting tag structure, so as to reduce the comparisons between element fan-outs and threshold. Finally, an optimized fragmentation policy was presented to avoid trivial pieces by binding XML sub-trees according to the ratio of element fan-outs and threshold. Our performance study shows that the document fragmentation algorithms perform well on execution time, granularity and other metrics.