الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 122 |  |  |
| | | from | 122 | | |
Abstract
The rapid and continuous growth of geospatial data generated from devices such as smartphones, satellites, and other IoT devices makes traditional GIS cannot support such a big data. Extracting knowledge from such a big geospatial data has become an extensive challenge. The traditional geographical information systems (GISs) lacks the adaptability of basic incorporated frameworks (e.g., local files frameworks and spatial database management systems (SDBMS)). Therefore, utilizing geographical information systems with cloud computing represents a new trend toward the progression of geospatial big data storing, processing, and its applications for GISs. Cloud computing represents the largest Information Technology (IT) transformation. Migration to the cloud become a mandatory demand. Doubtlessly, Cloud computing is expected to be the area of the most substantial growth and the most significant development. Recently, Hadoop is the most well known open source cloud-computing platform. Hadoop provides a solution for a huge data processing in many fields. Hadoop employs MapReduce to produce an efficient data processing framework. MapReduce was presented to give an effective distributed parallel processing paradigm with a fault tolerance and high scalability mechanisms. However, Hadoop has some deficiencies in terms of effectiveness, especially when dealing with geospatial data, a primary inadequacy is the absence of any indexing mechanism that could ease a specific access to spatial information particular areas, which consequently demand effective query processing. Because of this, an expansion of Hadoop, called SpatialHadoop, is developed. SpatialHadoop is a Hadoop framework supporting spatial information handling in light of MapReduce programming. A huge number of studies leads to that SpatialHadoop outperforms the traditional Hadoop in both overseeing and handling spatial data operations. SpatialHadoop provides an efficient query processing algorithms that access just a particular area of the information and give right query result. SpatialHadoop utilizes spatial indexes inside Hadoop Distributed File System as a method for spatial information efficient recovery. Indexing is the key point of SpatialHadoop better execution than Hadoop and the other systems.In our study, the proposed 4D and 2D PR-Tree are presented as a novel partitioning and indexing techniques in SpatialHadoop. In addition, a 4D point is created and developed as a new spatial data type in SpatialHadoop, which is necessary to implement and develop the 4DPR-Tree partitioning technique. An extensive experimental study has been performed to compare the proposed techniques with the state-of-arts SpatialHadoop techniques. Various techniques have been experimentally evaluated through different type of datasets (synthetic and real) with different distributions (uniformly, non-uniformly distributed data). 4DPR-Tree, 2DPR-Tree, and the state-of-arts SpatialHadoop techniques have been evaluated in several scenarios using a different set of spatial range and k-Nearest-Neighbor (kNN) queries. The experimental results have demonstrated the efficiency and accuracy of the proposed techniques.