Introduction to cloud computing
Cloud computing users do not need to manage or control the underlying infrastructure, including network, servers, operating systems, storage, and simply control the deployment of applications and environment.
Provide application service support platform and development platform, reduce the development cost of cloud computing application, and enhance the compatibility of application, and shield the complex underlying process flow.
Computing cloud platform consists of three layers:
Integration and exchange layer
This layer is mainly to solve the problem of integration between the system deployed on the cloud platform, based on the industry's mainstream SOA architecture.
Application running environment layer
This layer mainly provides the application system running, and the application system is based on the operating environment. This level also provides support for the deployment of the application system monitoring, diagnosis, data performance improvement and so on.
Storage and shared layer
This layer mainly focuses on the storage, integration, analysis and management of structured data and unstructured data (including large data).
The key technologies and functions of cloud computing are provided by Shaanxi Electric Communication Technology Co., ltd.. For more information, please visit: http://www.uetx.cc/
Resource scheduling
Resource scheduling technology is the computing cloud platform can automatically detect the application load, adjust the allocation of resources and scalability, in order to achieve load balancing and improve resource utilization, so as to protect the application of high scalability, so as to ensure the smooth expansion of application processing capabilities.
According to the different factors of platform size and application type, the emphasis of resource scheduling algorithm is also different. For example, when the scale of the platform is relatively small, the scheduling algorithm should focus on resource utilization; when the platform scale is large, the scheduling algorithm should pay more attention to the energy consumption of the cluster.
Parallel computing
For a specific task, by distributing data to different nodes, such as MapReduce based Hadoop can be processed in parallel (parallel) on the nodes of the data, which makes processing speed is very fast. Video file index, text information and image information are divided into several data blocks, which are stored in a redundant way. The client can read files from multiple DataNode and ensure high throughput data access.
Virtualization Technology
Virtual technology can improve the utilization rate of resources, and can carry out the resource deployment in accordance with the changes of user's business needs. In computing cloud environment, the virtual machine can run multiple virtual machines simultaneously through the physical hosts. Multiple virtual machines are running on the virtual platform, and the monitoring of multiple virtual machine operating system is realized by the virtual platform. Virtualization platform is the three layer structure, the lowest layer is the virtualization layer, provides basic virtualization capabilities, the middle layer is the control layer, provides the ability to perform the control functions, the top layer is the management, the implementation of strategic management, control, to provide a unified management of virtual platform.
Service performance
Ping An City to build the calculation of the cloud is mainly to complete the relevant video intelligence analysis tasks, intelligent analysis of the scheduling management and computing resources. Such as the need for complete path testing, legacy detection, tripwire detection, theft detection, loitering, sudden acceleration, retention of detection, suddenly appeared, face recognition, behavior characteristics, counting, population density, gun ball linkage, the intelligent, virtual bayonet, feature matching, feature recognition and color recognition, vehicle identification, human recognition and speed detection 22 intelligent analysis of the business. These twenty-two intelligent analysis services can be divided into three categories: real time intelligence analysis, video intelligence analysis and image quality detection.
Real-time intelligence analysis refers to the path detection, legacy detection, tripwire detection, theft detection, loitering, sudden acceleration, retention of detection, suddenly appeared, and intelligent video analysis is refers to the face recognition, intelligent abstract, vehicle identification and color recognition, image quality detection need sharpness detection, video noise detection, sharpness detection, anomaly detection of brightness, partial color detection, PTZ control function diagnosis, video freezing detection, video jitter detection, artificially blocked picture, video loss detection of ten kinds of image quality problems.
Computing Cloud Security
Cloud computing cloud must use a unified identity and security infrastructure to provide flexible security policy in the entire cloud, the user, security incidents and other centralized monitoring management and audit analysis, through efficient professional support platform and advanced monitoring tool for timely security incidents, timely grasp of security incidents, to calculate cloud environment for the early warning, emergency response and incident investigation to provide support, and to take active measures to protect user data, can meet the security needs of cloud computing cloud Center for comprehensive security.
Basic security is a basic protection layer, the main use of the firewall, vulnerability scanning, network encryption and other technologies to achieve network security. In Ping An City, cloud computing cloud deployments in the video, the calculation of cloud based security measures can rely on video