Querying Streaming XML Big Data with Multiple Filters on Cloud
Chi Yang, Chang Liu, Xuyun Zhang, Surya Nepal, Jinjun Chen · 2013
Nowadays, we enter a new era of data explosion which introduces the new problems for big data processing. Current methods for querying streaming XML big data are mostly based on events filtering techniques. It is well known that during the filtering, some data items have to be buffered before the filter can make the proper decision for adopting strategies to deal with them. Furthermore, for a single filter system, the buffer size often increases exponentially in the real application. Cloud is an ideal platform for big XML data processing with its massive storage and powerful computation capability. In this paper, we propose a new multi-filters strategy for querying streaming XML big data on Cloud. We show that the proposed multi-filters strategy can effectively share and reduce the filtering space and time consumption by fully exploit the scalability of Cloud. Furthermore, by deploying our multi-filters collaboration technique, the querying systems together can break the limitation of the theoretic concurrency lower bound. The empirical study shown in this paper demonstrates that our multi-filters strategy outperforms the single filter querying significantly.