Broadcasting and querying multi-dimensional index trees in a multi-channel environment

  • Authors:

  • Susanne Hambrusch;Chuan-Ming Liu;Sunil Prabhakar

  • Affiliations:

  • Department of Computer Sciences, Purdue University, West Lafayette, IN;Department of Computer Science and Information Engineeriny, National Taipei University of Technology, Taipei, Taiwan;Department of Computer Sciences, Purdue University, West Lafayette, IN

  • Venue:
  • Information Systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

The continuous broadcast of data together with an index structure is an effective way of disseminating data in a wireless, mobile environment. The availability of an index allows a reduction in the tuning time and thus leads to lower power consumption for a mobile client. This paper considers scheduling index trees in multiple channel environments in which a mobile client can tune into a specified channel at one time instance. Let T be an n-node index tree of height h representing multi-dimensional index structure to be broadcast in a c-channel environment. We describe two algorithms generating broadcast schedules that differ in the worst-case performance experienced by a client executing a general query. A general query is a query which results in an arbitrary traversal of the index tree, compared to a simple query in which a single path is traversed. Our first algorithm schedules any tree using minimum cycle length and it executes a simple query within one cycle. However, a general query may require O(hc) cycles and thus result in a high latency. The second algorithm generates a schedule of minimum cycle length on which a general query takes at most O(c) cycles. For some queries this is the best possible latency.