Internet-scale information monitoring: a continual query approach

摘要: Information monitoring systems are publish-subscribe that continuously track information changes and notify users (or programs acting on behalf of humans) relevant updates according to specified thresholds. Internet-scale presents a number new challenges. First, automated change detection is harder when sources autonomous performed asynchronously. Second, source heterogeneity makes the problem modelling representing than ever. Third, efficient scalable mechanisms needed handle large growing thousands or even millions triggers fired at multiple sources. In this dissertation, we model users' requests using continual queries (CQs) present suite solutions scale over structured semistructured data A CQ standing query monitors for interesting events (triggers) notifies meet In first system level facilities building an system, including design development two operational OpenCQ WebCQ, engineering issues involved, our solutions. We then describe research challenges specific large-scale techniques developed in context WebCQ address these Example include how efficiently process queries, what effective distributed trigger capable handling tens firing hundreds sources, effectively disseminate fresh right time. have optimize processing grouping scheme, auxiliary structure support group-based indexing CQs, differential evaluation algorithm (DRA). The third contribution experimental testbed validate engaged both measurements real (OpenCQ/WebCQ) simulation-based approach. To knowledge, documented dissertation date one focused study queries.