MobileWireless Middleware, Operating Systems, and Applications. Second International Conference, Mobilware 2009, Berlin, Germany, April 28-29, 2009 Proceedings

Research Article

Scalable Interactive Middleware Components for Ubiquitous Fashionable Computers

Download221 downloads
  • @INPROCEEDINGS{10.1007/978-3-642-01802-2_11,
        author={Gyudong Shim and Kyu Park},
        title={Scalable Interactive Middleware Components for Ubiquitous Fashionable Computers},
        proceedings={MobileWireless Middleware, Operating Systems, and Applications. Second International Conference, Mobilware 2009, Berlin, Germany, April 28-29, 2009 Proceedings},
        proceedings_a={MOBILWARE},
        year={2012},
        month={5},
        keywords={Ubiquitous middleware human machine interactions tuple spaces spatial query indexing},
        doi={10.1007/978-3-642-01802-2_11}
    }
    
  • Gyudong Shim
    Kyu Park
    Year: 2012
    Scalable Interactive Middleware Components for Ubiquitous Fashionable Computers
    MOBILWARE
    Springer
    DOI: 10.1007/978-3-642-01802-2_11
Gyudong Shim1,*, Kyu Park1,*
  • 1: KAIST
*Contact email: gdshim@core.kaist.ac.kr, kpark@ee.kaist.ac.kr

Abstract

The middleware for location based interactive applications requires scalability in large scale spaces. As the number of users and target services are increased, the server has to process massive spatial queries and event handling requests efficiently. Our middleware components are developed to extend the U-interactive system for large scale environments. The system manages the location information for large number of users and target objects. In addition the system handles events caused by user commands. We developed efficient tuple indexing and query mechanism by composite keys. As a new spatial query, Fan search is invented to provide efficient target selection by distance and angle. We optimized the query processing by efficient node traversing and data-aware interval skipping. The tuple matching process is performed in bounded time up to 100,000 objects. Fan search with C-Cuve has superior performance than Z-Curve in high density nodes in the experiment.