Research On Real-time Massive Data Processing For Personal Health Monitoring System Based On Wearable Devices And Mobile Internet

Cardiovascular and cerebrovascular diseases have become one of the main disease threatening human health.These diseases have the characteristics of sudden and concealment.When sudden cardiac arrest happens,the best rescue time is very short which is always shorter than 4 minutes.It is necessary to provide real-time health monitoring for patients.Due to the technical limitations,traditional monitoring mode always use off-line equipment to monitor the state of heart and brain in 24 hours.And the monitoring data need to be submitted to the hospital for analysis.The traditional monitoring is not in real time.With the development of wearable devices and mobile Internet technology,the realtime monitoring technology for cardiovascular and cerebrovascular diseases have a great progress.It makes the real-time healthy monitoring become possible.Remote health monitoring systems are mainly used in cardiovascular disease and major disease prognosis custody,which have the system requirements for real-time and a lot of users.The systems may have tens of thousands to hundreds of users.When a large number of users upload data,the system's storage systems need to dispose intensive I / O access requests,and continued accumulation of health data will challenges on storage capacity.According to the most typical cases of emergency processing time requirements,the massive health data need to be queried and disposed in real-time which brings higher requirements of real-time access and processing.This paper analyzes the characteristics of the practical application of remote health monitoring systems,designs the scheme of stream data processing,which improve real-time data storage,retrieval and processing capabilities.The main research work includes the following aspects:(1)Analyzes the characteristics of remote health monitoring systems and data processing,review the overall framework and operational processes of remote health monitoring system.According to the health care system's stream data processing features,designs hierarchical stream real-time data storage and processing solutions,and build a prototype system.(2)Healthy stream real-time data updating frequently which cause greater maintenance cost of the index structure.And the keyword-based query performance of key-value database is bad.This paper proposes a keyword indexing strategy based on space Z Curve and R-tree indexes.And according to the practical application need of real-time system,optimize the indexing strategy further.Experiments show that the proposed indexing strategy can effectively reduce the size of the index space and index structure maintenance cost,reducing system response time on keyword-based queries.(3)In order to solve the storage capacity challenges,the system needs to migrate the historical data to the disk based database orderly.In order to reduce the impact on system performance by data migration,we use the MMDB(main memory database)Redis' s master-slave replication and persistence mechanism,dynamic choose the migration opportunity to migrate historical data to disk-based databases.And study how to make users transparent access to these two databases.