Need Help?

OceanStor 9000 Stores Video Big Data in Cloud

01-02-2016 00:00:00

Background

The city of Karamay China set the goal of building both a Safe City and Digital Karamay project. To consolidate scattered IT resources, Karamay planned to migrate its existing IT infrastructure to a cloud data center or newly built IT system for central and easy resource management.

The project was designed to improve:

  • eGovernment
  • Smart campus
  • Remote healthcare
  • eCommerce

Challenges

Karamay required advanced functions and services from its video surveillance system, including:

  • 4K Ultra-High-Definition (UHD)
  • Cloud computing
  • Big Data handling
  • Intelligent video analysis

The project replaced 1,500 existing Standard-Definition (SD) surveillance cameras with UHD cameras plus added 3,000 new cameras. The system was designed for:

  • Public security management
  • Traffic management
  • Emergency handling
  • Accident warnings
  • Safety supervision

Karamay also required the system to jointly operate with peripheral systems like alarm, access control, and broadcasting systems.

Three were 3 major performance problems the system needed to resolve.

Video data protection

Video data protection was unsatisfactory, sometimes losing key data. Karamay wanted a secure, stable, and reliable video surveillance cloud storage platform to handle follow-up tasks such as crime investigation.

Information silos

Karamay’s legacy video surveillance systems could not communicate and share resources with one another without performance bottlenecks. In addition, the platforms were made by different vendors, preventing resource sharing and central management.

Video analysis

The legacy surveillance system could not provide advanced video analysis, preventing fast data mining and intelligent Big Data analysis. The system could also not process concurrent video reads and writes.

Solution

Huawei designed a distributed cloud storage solution based on its OceanStor 9000 Big Data Storage System. The system provides a total capacity of 17 PB.

It works by the bottom-level cloud storage system aggregating video data and transferring it to a remote cloud storage device for disaster recovery. The upper-level cloud system collects video data from the bottom-level system through a dedicated video network.

It uses advanced Real-Time Protocol (RTP) camera + distributed Network Attached Storage (NAS) technology. Network communication protocols meet GB28181 standards for a more stable and open system architecture that is compatible with a wide range of front-end cameras and is easier to operate and maintain.

OceanStor 9000 video cloud storage solution

OceanStor 9000 video cloud storage solution

Direct video stream storage

Video data is directly written to storage devices to reduce video server purchase and deployment costs and improve system reliability.

High-Volume, real-time browsing

The system’s cloud platform architecture allows users to view 1,000 channels of HD video simultaneously.

Online data sharing

The dedicated video surveillance network, public security and eGovernment intranets, extranet, and Internet are connected and can communicate with one another, making the video surveillance data available to the government and public.

Data protection

  • SecureVideo: SecureVideo tocology protects normal recording and playback when more than one disk is faulty
  • File-level data protection: File-level data protection provides redundancy across hardware nodes. The failure of one or more physical nodes does not damage data system integrity
  • Disaster recovery: The system independently configures disaster recovery links with important cameras in sensitive areas. Video copies collected by these cameras are automatically synchronized with remote cloud storage systems for enhanced video data protection

Resource integration

OceanStor 9000 distributed storage provides an ultra-large video storage resource pool that delivers industry-leading scalability. Users can schedule video storage resources on demand to eliminate information silos. The system is highly integrated with multiple intelligent subsystems, including:

  • Video identification
  • Emergency scheduling
  • Transportation security
  • Warning analysis

Video analysis

Huawei’s video cloud storage system provides advanced technologies that improve the indexing efficiency of video images and supports a 4:1 read/write ratio. The technologies include:

  • Video stream integration
  • Intelligent pre-fetch
  • Straight through I/O

It can evenly allocate, dynamically manage and process thousands of concurrent video analysis requests from thousands of front-end HD cameras. During video analysis, a Hadoop Big Data platform interfaces with the cloud storage system to evenly distribute tasks, automatically balance capacity and performance, and increase distributed processing efficiency.

The system correlates videos for management, including:

  • Original videos
  • Analysis process videos and images
  • Crime investigation videos and images

Distributed concurrent querying and collated analysis are provided to manage crime investigation information, helping users filter crime information and quickly solve cases.

Third-Party vendors

Huawei cooperated with video surveillance platform vendors to fully verify the functions of the cloud storage platform. Functions tested included:

  • Load balancing of stream media modules
  • Video downloading
  • Simultaneous playback of multiple videos
  • Support for IPv6
  • Stream forwarding reliability

Benefits

The Safe City video cloud storage solution designed for Karamay has several benefits.

Reliability

Video data and disaster recovery technologies ensure zero loss of key video data.

Convergence

The Big Data video storage platform consolidates the video resources of independent systems, allowing the resources to be centrally scheduled.

Efficiency

Up to 7,000 hotspot videos can be concurrently accessed, facilitating decision making and improving the emergency handling efficiency by 50%.