parallel programming

The era of big data is approaching with the rapid growth in development of complex information and communication technologies like internet and 3rd and 4th generation of Mobile Networks (3G/4G). These days more than ever, real-time and concurrent access to heterogeneous data from different sources and with different formats is made possible. Meanwhile, with recent advances in sensor technology and in order to monitor, explore and visualize complex spatial systems, a large amount of data in different Spatio-temporal scales is being generated every day. For example, according to the estimation by United Nations Initiative on Global Geospatial Information Management (UN-GGIM), 2.5 quintillion bytes of data are generated every day which large portion of the data is location-aware. This unprecedented trend of spatial data generation provides new opportunities for information and knowledge extraction, industrial development and business decision making. Although the big data brings new chances for scientific, business and engineering fields, it presents some challenges. To be more specific, storage, management, process and analysis of the spatial big data on traditional spatial information platforms is difficult and expensive. Such challenges affect modeling, analysis, simulation and concurrent access to spatial data. The need of real-time analysis in some applications like dynamic routing, fleet management or emergency management is also influenced by such limitations. In order to face spatial big data challenges, cloud computing technology to support spatial information applications appears to be very promising. Emergence of cloud computing technology provides an effective, scalable and affordable solution to big data challenges in spatial information application. Cloud computing provides fundamental support to address the challenges with shared computing resources including computing, storage, networking and analytical software. In this paper, we discussed fundamental theories and key technologies of cloud computing in storage, process and analysis of spatial big data. We have surveyed storage and management of big data using distributed file systems and NoSQL databases and made a comparison between different types of this databases. We also discussed recent trends and methods in parallel processing and big data analysis. MapReduce as a prominent parallel programming method and Hadoop as the most popular implementation of MapReduce are surveyed. We reviewed and made a comparison of spatial tools which is developed on cloud platforms. One of the most important challenges in spatial cloud computing that geospatial scientists is facing is spatial indexing and query processing. Due to distributed systems limits, developing dedicated spatial indexing and query processing techniques is needed. So we focused on novel spatial indexing methods and query processing technologies. As the case study, this paper surveys usage of cloud computing technologies in transportation, traffic and Intelligent Transportation Systems (ITS) and remote sensing and earth observations. The aim of this paper is reviewing and introducing fundamental theories, new technologies and recent trends of spatial big data to researchers of geospatial sciences.