XMind

A Distributed Spatial Index for Error-Prone Wireless Data Broadcast

1. Introduction

   1. organizes data objects in a certain linear order

   2. has a fully distributed structrue

   3. provides multiple search paths for same objects and shares links on common search paths

   4. very resilent to the inherent error-prone communication environment in wireless data broadcast

   5. efficiently supports a variety of location-based queries

2. Preliminaries

   1. system model

      1. performance metrics

         1. access latency

         2. tuning time

   2. conventional spatial query processing

      1. design requirements for air indexes

         1. the size of air index needs to be small

         2. air index can only be accessed sequentially

         3. query processing based on wireless data broadcase should start as soon as possible

         4. query processing should be resilent to the error-prone wireless communication environment

   3. air indexing

      1. conserving the energy of moblie clients

   4. procesing of location-based queries

      1. window query

      2. k nearest neighbor query

   5. hilbert curve

      1. 为什么偏偏选 hilbert 曲线，而不选其他的空间填充曲线

3. Distributed Spatial Index

   1. tree-based indexes are not appropriate for supporting spatial queries in data broadcast environment

      1. clients have to wait for the root node

      2. wireless broadcast is error-prone

      3. back tracking

   2. the index structure

   3. energy efficient forwarding

   4. error resilience

   5. multiple data types

      1. sequential approach

      2. integrated approach

4. Spatial Queries Processing

   1. snapshot query processing on air

      1. window queries

      2. k nearest neighbor queries

         1. conservation approach

         2. aggressive approach

   2. continuous query processing on air

      1. continuous window queries

      2. continuous nearest neighbor queries

5. Performance Evaluation

   1. point queries

   2. snapshot queries

      1. snapshot window queries

      2. snapshot K nearest neighbor queries

   3. continuous queries

      1. continuous window queries

      2. continuous NN queries

   4. computational cost

   5. power consumption

   6. error resilience

   7. multiple data types

1. 具体的算法不是很明白