What Are the Types and Requirements of Video Surveillance Storage

We first understand the types of network monitoring storage devices at this stage and their application characteristics in network monitoring.

Storage devices come up with different solutions according to different organization modes and network nodes, and can be roughly divided into digital video recorder (DVR), direct attached storage device (DAS) and network attached storage device (NAS).

Digital video recorder (DVR)

Stand alone digital video recorder DVR is the simplest and most economical way to operate high-quality digital element video monitoring system. It is not only slim, but also very lightweight. It can work normally by adding a camera.

DVR is most suitable for complex network monitoring solutions that do not need or have resources to manage. Because the DVR system is usually easy to operate and use, the on-site operators can easily view the video stored at any time, as well as carry out routine maintenance and data management. In addition, it is also the device with the lowest node restrictions on network installation.

Direct attached storage (DAS)

As the name indicates, DAS links the storage device directly to DVR or nDVR. It can be attached internally (the disk drive is loaded inside the DVR), or connected externally through a variety of raid external array storage interconnects and SCSI, USB or eSATA adapter cards. Although the cost is high and the RAID disk array is cumbersome and large, the storage capacity and expansion capacity provided by the external Das are only up-to-date. At the same time, the external Das enclosure also provides more disk drive slots than the DVR built-in hard disk slots. If more disk drives are needed, the external Das enclosure can increase the available storage capacity in a serial manner. It should be noted that the network setting node of this storage mode can also allow users to access video recording through the network TCP / IP architecture.

Network attached storage (NAS)

NAS looks like a mini network server. It generally helps to store files in the Ethernet network through TCP / IP, and uses the commonly used NFS or CIFS file level communication protocol. NAS deployment and maintenance adopt common IP and Ethernet communication protocols, and only a few IT personnel can manage NAS. Like the chassis of external DAS, NAS contains multiple disk drive slots, allowing users to choose the raid optical drive configuration that can best maintain a balance between performance, capacity and fault tolerance. Even if more capacity is required, NAS can be realized by adding external storage device modules. Objectively speaking, this kind of equipment is particularly flexible in the arrangement of network nodes.

What are the requirements for monitoring storage in the HD era?

In recent years, with the rapid development of video technology, the development of security video surveillance has also entered the high-tech stage. With the advent of the big data era, video surveillance technology and market have undergone qualitative changes. With the continuous improvement of security products and technologies, all walks of life have higher and higher requirements for video surveillance image quality, and then higher and higher requirements for data storage, which will inevitably promote the reform of video surveillance storage market.

From the application level, in fact, among all the monitoring image information, only 1% or less is useful, and 99% of the information is useless. But it is the 99% useless information that precipitates 1% useful information. Therefore, video information, whether useful or useless, should be stored, which is bound to bring huge storage pressure. Video surveillance images are often important evidence for post verification in many industry events and cases. In social security monitoring, banking, judicial and other system applications, the storage time of video surveillance images is often 1-3 months or even longer. The improvement of monitoring image definition will inevitably bring greater pressure on image storage.

First, the demand for mass storage. The monitoring data is 7 × 24-hour write, storage time from 7 days, 15 days... Even to 1 year, and the amount of data increases linearly with time. For chain commercial organizations, it is common for each branch to arrange 5 to 10 monitoring channels, which is difficult for the traditional storage industry and monitoring industry.

Secondly, the performance requirements. Video monitoring mainly refers to the writing of video code stream. The performance is characterized by how many code streams can be supported by storage (usually 384kbps-2mbps). In the case of multiple concurrent writes, it has a great impact on bandwidth, data capacity and cache, and has a great pressure on storage. At this time, storage needs to be optimized specifically for video performance.

Third, price sensitivity. The mass storage in the security monitoring industry has a large total capacity, resulting in an increase in the total price and cost. On the contrary, the cost per unit capacity (price per TB) is very high.

Fourth, the requirements of centralized management. In the application of security monitoring, due to the need for a large number of storage devices, the massive data in the storage devices must be effectively managed to provide convenient, reliable and transparent support for the use of end users.

Finally, networking requirements. TCP / IP network is the network foundation for the development of security monitoring technology in the direction of networking. The storage technology based on TCP / IP network will play an irreplaceable role in the networking process of security monitoring technology.

Enterprise monitoring storage focuses on cost saving.

With the popularization of video surveillance applications, enterprises are one of the main objects of video surveillance. Compared with home monitoring and public monitoring, enterprise monitoring has higher requirements for storage cost. On the premise of ensuring the integrity of the monitoring system and good storage effect, we should also take into account the cost.

In its article "the challenge of realizing low-cost enterprise class cloud storage services", CTO David flover of wikibon company believes that about 80% of applications have the possibility of reducing relevant storage costs by about 50%. This bloated "poo belly" is composed of a few Tier-1 and all tier-2 applications. We should simplify its architecture and management mechanism in the design idea. The article emphasizes that not every storage service needs the support of expensive service projects and management processes for low latency databases and applications. In fact, what many applications really need is to reduce the relevant storage costs as much as possible.

At present, hybrid arrays have found the best compromise between performance and cost, floyer wrote. Full flash arrays are still too expensive and focus on the unit IOPs cost of high-performance applications rather than the unit storage cost. As a solution combining the advantages of software defined storage and direct attached storage, server San brings us a more flexible and high-performance storage architecture, but its maturity is still insufficient - floyer believes that this situation will not improve until 2015 or later.

Most of the hybrid systems launched by suppliers can work, but he focuses on IBM's XIV. It can provide a simple way to operate, eliminate the need to tier the storage of target applications, and easily solve the problem of "proximity impact" (referring to those applications that require very high infrastructure resources, and their operation may crowd out other applications running on the same storage array) - more importantly, It only needs a small technical team to take care of it easily. He suggested that XIV be included in the RFP for shared storage services - although not perfect, he also made a detailed analysis of its specific advantages and disadvantages.

Floyer designed a set of comparison models, assuming that enterprise users have an annual revenue of US $10 billion and an overall IT budget of US $150 million. On this basis, he calculates the difference in use cost between traditional tiered storage and cloud optimized hybrid storage. According to the calculation of the model, after weakening the bloated state of the application, the overall cost of storage can be reduced from the original $16.169000 to $8.474000.

Five ways to redefine the storage architecture are more economical.

More efficient storage systems can free up staff time, benefit enterprises in other aspects, reduce costs, and accelerate the responsiveness to new businesses.

Organizations can adopt innovative technologies to redefine the efficiency and cost-effectiveness of their data centers through minor changes / enhancements. According to Dell's storage technology experts, there are five things that organizations should consider before choosing a storage architecture:

1. Utilize flash products at disk prices

Flash storage can guarantee higher performance and greater enterprise productivity. It uses a different way from disk storage, information retrieval is much faster, and has minimal susceptibility to damage. Flash memory storage has been very popular in the consumer market. Now it is becoming more and more popular in the enterprise market because of its potential in high-performance computing. When assembled in an efficient way, enterprises can take advantage of the benefits of flash storage at the price of disk.

2. Scalability and enterprise capabilities

When choosing your storage infrastructure, it is important to consider the growth of the company. Your data and storage requirements are bound to increase over time. Your storage architecture should be flexible to support new solutions without having to go through constant and expensive refresh cycles. In addition, it is important that the storage architecture you deploy has the characteristics that drive enterprise level decisions and can provide the elasticity, reliability and durability of enterprise level storage architecture.

3. Automatic layering

Automatic tiering means that data is stored only at the most appropriate level, optimizing performance and cost. The media selection corresponding to different data is suitable for their actual needs and budget considerations. The use of intelligent automation makes storage easy to manage and reduces storage complexity and storage cost. It is the symbol of future storage infrastructure.

4. Data deduplication and storage virtualization

Data deduplication technology refers to the process of finding and eliminating duplicate blocks in the data storage of different data sets, and the data storage requirements can be reduced by up to 90%. In addition, storage virtualization enables enterprises to abstract multiple storage devices into a single storage pool, thus helping administrators move data in the storage tier. Considering such technologies for your storage infrastructure can ensure greater benefits.

5. Integrated infrastructure

Converged infrastructure (CI) is such a data center management method that converts decentralized data center components, such as storage, server, network, etc., into a single device that can be managed centrally. It helps minimize compatibility issues, simplifies the management of servers, storage systems, and network devices, while reducing the cost of cabling, heat dissipation, power consumption, and floor space. Forrester's research found that 95% of it leaders and storage administrators said they saw the value of purchasing servers, networks, system management and storage from the same it vendor because they felt that a deeper relationship with fewer vendors could build trust, provide manageable technology synergies and simplify support