The industry-specific cloud is based on the integration of cloud computing, big data, intelligent algorithms and other core technologies, as well as the core concept of "cloud-edge hybrid computing and open compatibility", and aims to provide a video cloud solution based on IoT concepts in industries such as public security, transportation, energy, justice, culture, education, finance, and others.
The industry-specific cloud is made up of a three-tier architecture: "terminal, cloud, application". The intelligent terminal not only includes smart terminal equipment for collecting videos and images, but also includes intelligent storage and computing service equipment, which can extract and analyse image details such as human, vehicles and other objects in real time to provide image structured data needed for big data analysis. The integration of video data and various types of object-sensing data, combined with various types of intelligent algorithms, network storage resources, and intelligent edge resources, creates a unified network system. This system combines a data resource pool, an algorithm pool, and a device resource pool to provide open video services, smart services, big data services, basic security services, and cloud computing services. These services can be applied to meet the needs of different industries and add value to businesses.
The industry-specific cloud architecture and principles are in line with the core needs of industry users, technology development trends, and principles of sustainable development, and will satisfy industry digitisation and information trends in the coming years.
（1）More intelligent: using high-performance GPU chips, advanced deep-learning model, and massive data samples, the deep-learning algorithms can be trained to gain intelligent perception to more accurately extract unique identifying features of human, vehicles, objects and other data from videos.
（2）More open: supports algorithm database and is compatible with smart algorithms of many manufacturers. It supports GB28181, ONVIF, EHome, MQTT, EZVIZ, and other protocols, and is compatible with front-end equipment from mainstream manufacturers. It supports open cloud computing services, video services, smart services, big data services, security services, and provides a unified open interface. It also supports unified management and scheduling of cloud and edge resources.
（3）More reliable: where computing pressure has been decentralised, system malfunctions have also been decentralised; any central or network failure will not affect front-end video analyses and storage. Virtualisation, clustering, distributed, multiple copies and other technologies ensure the high reliability of calculation, storage, and data.
（4）More cost-effective: edge computing is more effective in reducing costs. First, compared to centralised equipment, edge equipment achieves faster cost reductions. Second, after reducing calculation pressure, edge equipment greatly reduces the need for centralised equipment, saving space and energy. Finally, the intelligent data generated by smart edge technology requires less bandwidth, saving network traffic.
（5）More practical: edge computing can reduce analysis latency and enhance business agility; edge node analyses and computing results are fed back to the video capture system, where image quality and encoding are optimised. Cloud computing architecture makes the system more scalable and compatible, and is therefore more suitable for users to build valuable applications.