Viewing the Development Trend of Modular Data Centers from *Policing Cloud Big Data

Silent data speaks

Case 1: The theft occurred in a college’s dormitory. Several computers and mobile phones were stolen. The local authorities immediately inquired about the surrounding video surveillance and peripheral bayonet videos. According to the time of committing the crime, the suspected vehicle passing time of the bayonet was calculated and locked quickly. Suspected vehicles, according to the license plate number, the police quickly locked the trace of the suspect. In the "big data combat platform", enter the license plate number, the vehicle's trajectory at a glance, and then collide with the "electronic fence" information that can capture the mobile phone signal, who was on the car after the incident, and soon there is result.

Case 2: A foreign license plate full of goods, imported from a high-speed exit into the urban roads, the drivers are eager and swift, ignoring the traffic warning signs that the trucks are forbidden to enter the elevated roads in the urban areas, and they slammed into a certain road. Elevated. Unexpectedly, only five or six kilometers of traffic, the vehicle was quickly stopped by *. According to the local * introduction, thanks to the intelligent video surveillance system, this car will illegally send illegal information to the nearest * police mobile phone within 10 seconds when it illegally enters the elevated portal. In this way, * can be law enforcement in time.

* Characteristics of police cloud big data

Today, the use of big data to help * successfully solve cases is too numerous to mention. The above are just two common cases. With the development of information technology, the information-led policing system has been promoted to help solve challenges easily. * A large number of policing data and social resources data are collected in the system database. The policing cloud is used as a carrier of big data and has the typical characteristics of big data:

The first is massive, large enough to "amount of data that cannot be managed with the current technology";

The second is diversity, and the types of data are complex, including not only traditional formatted data, but also web logs, videos, pictures, geographic location information, etc. from the Internet. Unstructured data accounts for 75%-95% of the total amount of stored data. These unstructured data cannot be processed with current technology.

The third is low value density and high commercial value. Getting data from large amounts of low-quality, low-value data is costly. Take video as an example. The data that may be useful during continuous and continuous monitoring is only one or two seconds.

The fourth is the speed of processing. The frequency and frequency of data generation are very fast and require real-time processing; this is fundamentally different from traditional data mining techniques.

Policing Cloud Big Data Challenges Infrastructure

With the deepening of the construction of the “policing cloud”, the number of information system data is increasing, and there is an explosive growth trend. The information systems built with the traditional construction model are becoming larger and more complex, and the pressure for management and operation and maintenance has also increased. The data center infrastructure is developed following the development of the upper layers. The traditional data center architecture pales in comparison with the data center construction requirements in the era of policing cloud big data.

First, the power density is low: the traditional data engine room is usually in the form of air supply, supporting the power density is low, in the application of high-density servers, prone to local hot spots, resulting in equipment failure. Therefore, the traditional air distribution system is no longer applicable to the new data center for policing cloud big data.

2. Inefficient operation and maintenance: There are 6 ICT operators in a provincial office, who need to be responsible for ICT, video, system security, and informatization project management. Only 1 person is responsible for the operation and maintenance of the computer room infrastructure, and can only pass through. The manual patrol method has very low efficiency, even equipment failures, and it only took half a day for services to respond to the situation.

III. Difficulties in Capacity Expansion: Traditional data houses have been solidified since the beginning of construction. It is very difficult to achieve future capacity expansion. For data houses that need to be expanded, plans must be put forward at the beginning of design to embed the necessary infrastructure at one time. .

New modular architecture emerged

In this context, if a modular architecture is adopted, the above problems can be solved. The Modular Data Center is mainly targeted at data centers in the city bureaus, districts and counties, and police stations in Ping'an City to provide them with the hot water and power needed for data center infrastructure. The program integrates power supply and distribution systems, row-level cooling systems, cabinet systems, monitoring systems, and integrated wiring systems. It adopts a modular architecture, factory prefabrication, and rapid deployment to help build a reliable, green, and intelligent data center.

First, high power density: Modular data center adopts row-level air-conditioning approach to heat source, which can achieve higher power density than traditional refrigeration systems, and because of the decentralized arrangement characteristics of row-level air conditioners, it can support different implementations in different micro-modules. The demand for power density more meets the demand for power density that is different for different services.

Second, to simplify the operation and maintenance: Modular data center has a unified operation and maintenance interface, interface interface is easier to manage, by a unified modular equipment provider is responsible for solving all the professional connectivity problems involved in the infrastructure, fundamentally eliminate the different professional Pushing each other happens. Second, the modular data center has intelligent features, such as battery capacitor fan warning, battery shutdown and other functions, making operation and maintenance easier.

Third, the flexible expansion: modular data center to achieve pre-integration of the various subsystems in the factory, greatly simplifies the installation of the project, reduce on-site deployment time, and to achieve the flexibility of the architecture does not need to determine the entire room as the traditional machine room After planning, the infrastructure will be unified. Modularized data centers can be prefabricated along with project construction, planned according to business needs, and deployed step by step in stages. Similar to the way of building blocks, they can be expanded with the needs of the system.

Figure 1 Application of Medium-sized Modularized Data Centers in the Computer Room of Provincial Offices and Municipal Offices

Figure 2 Application of Small Modular Data Centers in District and County Computer Rooms

Figure 3 Application of Micro Modularization in the Police Station Room

The modular data center is not only applicable to the policing cloud application of the industry*, but also applicable to other industries such as finance, transportation, and energy. The latest ICTresearch report shows that the modular data center market is increasing year by year. In 2015, Huawei Modular Data Center led the industry with 17% market share. (Figure 4 below)

Figure 4 Market share of each brand in China's modular data center market in 2015 (Source: ICTresearch 2016)

The modular data center's features of practicability, intelligence, and flexibility are more suited to the needs of today's big data. High-density, low-power data centers will become the mainstream. It is believed that the modular data center will gradually replace the traditional construction model and protect the data centers of all walks of life.