Inspur Storage and its partner Tiandi Weiye jointly launched an integrated smart traffic checkpoint solution for traffic security (vehicle and personnel checkpoint monitoring). Based on this joint solution, the public security and traffic management industry can implement AI cluster analysis and data mining applications, find vehicles and recognize faces in real time from massive video data, and build smart security.

During the IPF2018 Inspur Cloud Data Center National Partner Conference, Inspur Storage and its partner Tianjin Tiandi Weiye Digital Technology Co., Ltd. (hereinafter referred to as "Tiandi Weiye") jointly launched integrated wisdom for traffic security (vehicle/personnel checkpoint monitoring) Traffic stop solution.

Through this solution, security and traffic control personnel can use fast face detection technology to find faces from video resources in real time after extracting face pictures to lock people and achieve rapid identity search.

At the same time, for event retrieval, you can set tripwires and perimeter rules on the video screen to retrieve the targets that trigger these events. The retrieved tasks can be sorted and displayed according to target attributes. Double-click the snapshot to locate and play the original video of the snapshot. When a suspect target is found, the suspect video or picture can be turned into an investigation clue.

In addition, using fast vehicle detection technology, vehicles can be searched in real time from video resources, and the license plate can be output while identifying the suspect's car to achieve rapid vehicle search.

Inspur Storage and Tiandi Weiye launch a joint security solution that can quickly detect vehicles and recognize faces

How to find the needle in the haystack with massive security data?

Tiandi Weiye is a well-known enterprise in the security field, a leading industry solution expert, and a professional digital security product R&D and production base in China. Inspur is China's leading cloud computing and big data service provider, and has formed overall solution service capabilities covering the three levels of IaaS, PaaS and SaaS. Inspur and Tiandi Weiye started cooperating in 2009. On May 15 last year, Tiandi Weiye and Inspur Group further signed a strategic cooperation agreement, laying a solid foundation for both parties to carry out in-depth cooperation in the field of security.

Tiandi Weiye official website

Starting from the first batch of 50 demonstration cities for video surveillance construction and networking application in 2016, China has incorporated the construction of video surveillance systems into the "13th Five-Year Plan". This security system, from cameras to storage systems, is a network with a huge amount of data, which brings huge challenges to traditional security.

Take the traffic control business needs of a standard urban district and county as an example. Generally speaking, the public security and traffic control departments will plan and build four major platforms to collect and upload a large amount of illegal information. The four major platforms will jointly process services such as ultra-low speed violation inquiries, dynamic violation captures, license plate violation inquiries, and vehicle process inquiries.

Building a district-level information sharing service platform requires the integration of data resources from approximately 50,000 videos, 50,000 lanes, and 2,500 WIFI collection systems. Various data resources such as district-wide videos, pictures, vehicle structures, WIFI structures, RFID structures, etc. need a data storage system to carry them.

Traditional security application issues

First, passive application.Traditional passive applications need to deal with dozens or even hundreds of videos at the same time, making it difficult to detect sudden and abnormal events in the video footage in a timely manner, and the video verification process requires a lot of manpower and time.

Secondly, the data cannot be used effectively.Massive video data cannot be effectively mined and analyzed in depth, which not only makes the monitoring system run inefficiently, but also causes a huge waste of video equipment and data resources.

Again, storage pressure is huge.Massive unstructured data, mostly useless data, puts great pressure on storage devices. It not only wastes storage space and increases storage costs, but is also not conducive to rapid data retrieval.

SDS as the core, serving smart security

The SDS (software-defined storage) solution for security scenarios provided by Inspur Storage and Tiandi Weiye is precisely to help users meet the above challenges.

First, the system achieves linear expansion of storage performance.ICFS supported by Inspur Software-Defined Storage AS13000 is a software-defined storage file system developed by Inspur Innovation. It adopts decentralized technologies such as metadata clustering and inter-node data redundancy to allow the system's data and metadata to be distributed on each node. This eliminates performance bottlenecks and allows storage capacity and computing power to increase linearly as the number of nodes increases. ICFS also provides a globally consistent cache, allowing the cache capacity to grow linearly with the increase of nodes. At the same time, the capacity of the global cache also increases, and increases the hit rate of hotspot data, greatly reducing hard disk I/O access. frequency, improving overall system performance.

Secondly, it supports AI applications through ultra-high storage performance.As a software-defined storage with scale-out (horizontal expansion) architecture, AS13000 adopts small file IO aggregation, IO path optimization, network delay optimization, all-flash media adoption, network delay optimization (RDMA), metadata and data operation control, and directory classification. Film and other technologies achieve ultra-high performance to help AI cluster analysis applications in the video surveillance industry.

Again, the communication speed is increased by 4-6 times.Inspur software-defined storage AS13000 also achieves higher communication speed through the internal IB networking configuration scheme, reaching 4-6 times the previous rate; it supports RDMA (Remote Direct Memory Across), achieving low-latency and high-performance data transmission .

Finally, the business pressure can be fully balanced.Inspur software-defined storage AS13000 adopts an automatic load balancing solution, which makes the business pressure between each node in the cluster basically equal to support the most commonly used polling and balancing strategies based on the number of connections based on business needs.

Inspur software-defined storage AS13000

Serving the video surveillance industry

The scenario-based solution for the security industry based on Inspur's software-defined storage system AS13000 solves the challenges of storage capacity, performance, cost and reliability in traditional security systems. Inspur Storage and Tiandi Weiye work together to help the public security traffic management industry realize AI cluster analysis and data mining applications and promote the development of the security industry.

Taking face recognition as an example, through the practical application of public security face recognition for dynamic/static portrait data analysis application systems and video structured data analysis application systems, using deep learning algorithms, the recognition rate has exceeded the ability of the human eye, and can distinguish gender and age. , glasses, sunglasses, masks and other face attributes, it still has a high recognition rate even after wearing glasses, sunglasses, masks and other camouflage behaviors. The face rotation angle has a wide adaptability range, and the number of faces in the recognition screen exceeds 100.

"Inspur's ecological partner business is growing rapidly," said Sun Gang, general manager of Inspur's storage product department. "In 2017, the storage business of Inspur's partners has grown significantly, achieving breakthroughs in finance, energy, medical, radio and television, video surveillance and other industries." Higher growth. Inspur has cooperated extensively with ISV manufacturers and achieved breakthroughs in more than 200 customer projects. In the monitoring industry, Inspur Storage has strategically cooperated with Tiandi Weiye, focusing on areas such as intelligent transportation and safe cities, and launched joint solutions . This proves that Inspur Storage has strong market competitiveness and enhances Inspur Storage’s confidence in increasing the construction of intelligent ecosystem."