Geostream Computation Model And The Edge Computing Experiment

In recent year,the rapid development of mobile sensors and wireless networks produce a large number of moving spatial data.In addition to the temporal nature of this stream data,various sources provide stream data that has the geographical locations and/or spatial extents and is known as "GeoStream".The theory and method of traditional spatial database can hardly deal with the dynamic GeoStream.On one hand,this amount of streamed data has been a major propeller to advance the state of the art in geographic information systems.On the other hand,the ability to process,mine,and analyze that massive amount of data in a timely manner prevented researchers from making full use of the incoming stream data.This paper focus on the research of real-time spatio-temporal data stream processing.From the point of model representation,data organization,query optimization,we explore the steam data model of moving object based on signature algebra,design the spatial index with the high update rates as the synopsis data structure of spatio-temporal data stream and put forward a framework towards efficient real-time managing and monitoring of mobile objects through distributed spatiotemporal streams processing on large clusters.Specific contents are as follows:(1)The expression method of signature algebra and analysis the abstract data model type system and operator of moving object are expounded.We distinguish two different data types of online and offline data.We propose a method to upgrade the data model from offline to online spatio-temporal stream and finally establish spatio-temporal stream data model.(2)The moving characteristics of moving objects in the Euclidean space and road Network space respectively and design of two different discretization methods for abstract data model in the above two condition are studied.In order to adapt the high rate of time-stamped location updates,in the European Space,we add secondary index on traditional spatial index to complete the bottom-up fashion update and in the road network space,we design the node table,the edge table for moving objects and propose the corresponding CKNN algorithm on update message.Finally,the synopsis data structure of spatio-temporal data stream is implement.(3)Two big data paradigms of spatio-temporal data stream for the spatial query on single moving data set(knn and range query)and between two moving datasets(spatial join query)are proposed.We explores the factors that affect the output rate of spatio-temporal stream processing,which can also verify the feasibility of our approach.(4)In order to address the large network bandwidth consumption caused by geo-distributed on spatio-temporal stream data source,exploratory experiment of spatio-temporal stream processing with the help of new edge computing framework is try to conduct.Based on this framework,the edge cluster is deployed in four cities which incluide BeingJing,HangZhou,ZhenZhou and HuZhou.The experiment is conduct from the points of network transmission efficiency,edge cluster number and local,global task ratio to increase the spatdata to calculate the output rate of processing.Research and experiment results show that the theory and method of real-time spatial computation for spatio-temporal stream data,which is referenced by International advanced moving object data model,real-time moving spatial index,spatio-temporal stream processing on cloud and edge computing framework,can effectively improve the output rate of processing.