International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 99 - Number 18 |
Year of Publication: 2014 |
Authors: Navreet Kaur, Harwinder Singh Sohal |
10.5120/17473-8327 |
Navreet Kaur, Harwinder Singh Sohal . XML Parsing on Multicore Processors and Data Representation in .NET Tree Control. International Journal of Computer Applications. 99, 18 ( August 2014), 29-35. DOI=10.5120/17473-8327
The purpose of this research is to optimize the parsing process of the XML files. There are several ways to parse the XML files. But to comply with the advanced multicore CPUs and their fast performance the XML parsing logics need to be refined and optimized with parallel processing approach. The parallel XML parsing is a step towards this approach. It makes the reading of XML data faster because parser runs on more engines to extract the data. There are several advantages of parallel XML parsing like fast execution, high throughput, time saving, proper CPU utilization and load balancing. To perform the parsing processes simultaneously, the XML files need to be split in small uniform portions. Now it will execute the parsing logic on multiple threads on each CPU's core to parse the each portion of XML file without interfering with each others. In other words, an each segment will be an input to the parser running on different threads on different CPU cores. To enhance the system performance the multicore processors based devices have been introduced. Such system's processing is much faster than conventional sequential processing systems especially when it does repetitive calculations on vast amounts of data. This technique becomes more important when a candidate system or development application is model based application which operates on the XML files. This approach plays a significant role to enhance the application's capability to process large amount of data, improve application performances by providing quick results and eventually expeditious the application processing and dependent operations.