IntroductionXML is a semi-structural and standard document format to exchange data. Elements in XML documents are regular and there are structural relationships between them [1]. In fact, processing queries in XML should recognize these structural relationships. They also should determine the order of elements in a document. Node labeling in XML data is one way to increase the efficiency of query processing. Labeling means allocating a unique identifier to each node in XML documents [2]. A labeling scheme encompasses traversal or browsing the document, analyzing the elements, and assessing available relationships between elements. So, it should generate small enough labels in order to be processed efficiently both in initial label assigning as well as when queries are issued.A challenging problem of the existing labeling schemes is the need for relabeling nearly all the existing nodes after inserting new nodes in XML documents. While update in XML data is a usual operation in many real-world applications, e.g. stream data, relabeling will influence the query performance, especially when large-size labels are assigned Abstract Query processing based on labeling dynamic XML documents has gained more attention in the past several years. An efficient labeling scheme should provide small size labels keeping the simplicity of the exploited algorithm in order to avoid complex computations as well as retaining the readability of structural relationships between nodes. Moreover, for dynamic XML data, relabeling the nodes in XML updates should be avoided. However, the existing schemes lack the capability of supporting all of these requirements. In this paper, we propose a new labeling scheme which assigns variable-length labels to nodes in dynamic XML documents. Our method employs the FibLSS encoding scheme that exploits the properties of the Fibonacci sequence to provide variable-length node labels of appropriate size. In XML updating process, we add a new section only in the new node's label without relabeling the existing nodes while keeping the order of nodes as well as preserving the structural relationships. Our labeling method is scalable as it is not subject to overflow, and as the number of nodes to be labeled increases exponentially, the size of labels grows linearly, which makes it suitable for big datasets. It also has the best performance in computational processing costs compared to existing approaches. The results of the experiments confirm the advantages of our proposed method in comparison to state-of-the-art techniques.

No abstract

No abstract

No abstract

No abstract