A distributed selectivity-driven search strategy for semi-structured data over DHT-based networks

Distributed Hash Tables (DHTs) are widely used for indexing and locating many types of resources, including semi-structured data modeled as XML documents. A common distributed strategy to process an XML query over a DHT consists in splitting it into a set of simple path queries, and resolving each of them separately. The traffic generated by this strategy grows with the number of paths in the query. To overcome this drawback, an alternative strategy consists in resolving only the sub-query associated with the most selective path, and then submitting the original query to the nodes in the result set. A first goal of this paper is to provide an analytical and experimental study of the two strategies to assess their relative merits in different scenarios. On the basis of this study, we introduce an Adaptive Path Selection (APS) search technique that resolves an XML query in a distributed way by querying either the most selective path or the whole path set, based on the selectivity of the paths in the query. The effective use of APS requires that the querying nodes know in advance the selectivity of all the paths. Addressing this problem is another goal of the paper, which is achieved through: i) The definition of a space-efficient data structure, the Path Selectivity Table (PST), which given any path, returns an estimate of its selectivity. ii) The definition of an efficient strategy that builds the PST in a distributed way and propagates it to all nodes in the network with logarithmic performance bounds and without redundant messages. Experimental results show that the PST accurately estimates the path selectivity values, and that the traffic generated by the APS algorithm using PST-estimated selectivity values is comparable to that produced by APS assuming to know the real path selectivity values.