Large Data products and Gang

Talk about the big data market opportunities, there are enterprises from one machine into, there are enterprises from the data analysis solution cut, there are soft and hard. There is a point of view that the real big data solution is to produce results that have not previously been used. If it is unprecedented, there is no precedent for local companies to follow. And the reality is that many companies are claiming to be big data, so how do they do it? How is the gap with the leading program?

Software strength to pull the whole machine competitiveness

The local enterprise software ability is weaker, makes the hardware performance can be used well to become the difficult problem.

The concept of large data is hot, and many people in the industry hold the view that "the tipping point in the big data market is breaking out". In anticipation of the market temptation, local manufacturers have to light their own big data development roadmap.

Not long ago, good at the dawn of the server to launch its new products large data integration machine. According to Lichun, president of Shuguang Company, the All-in-one machine separates the data storage unit from the processing unit, and the storage system uses large-scale distributed technology. At the same time, the dawn also proposed its own large data integration solution, claiming to be based on its cloud computing strategy and the National Dawning Cloud Computing Center.

How close is the big data market and the hardware and the hot big data-all-in-one? From the hardware base, An Hui, director of the Institute of Software and Information services, told China Electronic News reporter, large data-related hardware mainly involved in bus, memory, CPU, display equipment. Large data requires the CPU's instruction set to be more suitable for unstructured, semi-structured data processing, faster memory, memory or the logical structure of the motherboard is more suitable for high-speed computing, such as memory calculation, hardware equipment inside the data bus bandwidth and hardware equipment network bandwidth wider, display devices can give users more intense immersion. At the same time, due to the rapid increase of data volume, the scale of parallel database increases, resulting in a sharp increase in costs. For cost reasons, an increasing number of enterprise-level users will be moving from high-end servers to a large cluster platform consisting of low-end hardware.

Domestic and foreign enterprises are in succession to launch large data integration machine, and manufacturers of high enthusiasm, the market reaction is not so eager. Pan Yonghua, a senior analyst at IDC, admitted in an interview with the China Electronic News reporter large data in the market is still relatively fresh things, there are not many applications need to use large data integration machine, on the contrary is based on the traditional relational database of one machine user market larger, become a mainstream market application. These database integration machines are mainly used in the analytical type, the user transaction performance requirements are higher applications.

It is because of large data integration machine market and soon, but also did not open markets, domestic and foreign manufacturers of the product gap is small. Pan Yonghua that foreign manufacturers accumulate a lot of data processing, domestic manufacturers mainly from the hardware point of view, in software accumulation and foreign manufacturers gap is large.

The strength gap in software is the main factor of the gap between the large data integration machine and the local large data machine. Pan Yonghua to the "China Electronic newspaper" reporter analysis, foreign enterprises to take "soft binding hard" strategy launched in one machine, is the software installed on a specific basis of hardware, to specific applications to tune, so that product performance, usability and manageability have a certain upgrade. The local enterprise's soft rib is the software ability, this makes the hardware the performance to be able to use very well to become the difficult problem. In addition, the local manufacturers push large data all-in-one, some are based on Hadoop open source platform developed, and open source platform in terms of availability, security, high reliability problems. is the commercialization of open source software a good solution? Is it true that the advantages of software hardware integration are reflected? These are the problems that domestic manufacturers need to consider.

The ability of product to be tested

Domestic manufacturers need to work on the product side, rather than simply bundling software tools into their own solutions to sell.

The large data industry chain covers 3 levels of hardware, software and services. From the market size, to service and hardware-oriented, software share is small. Large data service is the focus of many software vendors.

Zhou Xia, vice president of the company, said that making large data produce its big value should be the goal of each big data enterprise, and how to make the data valuable? It is the need to make the data into good information and generate value in it, which is also the core location of the same-side big data.

The research shows that the industry market's large data value density is the highest, after the shallow processing can produce great value, therefore, the industry users have become the most urgent demand for large data mining market target users. However, although the industry data already exist, the enormous value has reached a consensus, but because of the general lack of high-grade technical staff, lack of experience, many industries on the application of large data demand is not clear. "Most Chinese companies think big data is hot, but they are confused about what can be solved with big data," he said. Local enterprises need to find the right needs, find the right positioning. "Pan Yonghua said.

Large data solutions require sectoral considerations. An Hui told the "China Electronic newspaper" reporter, in the field of industry, different enterprises, business is not at the same time, due to the analysis of the data set and analysis of the target differences, the use of large data technology and large data systems may also have a considerable difference. is due to the specific application of close contact, or even one-to-one contact, only to make "application" become an indispensable part of the content of large data. At the same time, because of the need to use a number of general-purpose software and some basic analysis model, so if the solution can be layered, you can think that there is a common layer. But for each industry domain, enterprise's personalized software, model, then on this.

Pan Yonghua that the domestic manufacturers market opportunities are more for the vertical industry solutions, because they understand the needs of Chinese users, integration into large data solutions, to reflect the advantages. For the general tool market, domestic manufacturers can choose some narrow areas, make their own characteristics, improve the degree of product. Although the market is based on foreign products and open source software, it does not preclude the existence of opportunities for development. For example, the data visualization tool, there are entrepreneurial companies with the product in the United States successfully listed cases.

"Domestic manufacturers need to work on the product side rather than just bundling software tools into their own solutions to sell." Pan Yonghua stressed.

Start with the data center

Data centers or information centres that are eager to establish a data-resource system are large data partners for IT companies.

Government investment to stimulate the construction of intelligent city, many manufacturers also in the wisdom of the city's opportunities to benefit a lot. If the core of the intelligent city is what, then large data processing should be one of them. Large data help government departments or industry enterprises to make better decisions, improve service management level, and make city management smarter. Compared with the market of enterprise Big Data precision marketing oligopoly development trend, the government-led smarter City now seems to have more opportunities, which is why many large data makers are positioned in the government industry market.

For the smart city's big data market, where do companies start? Li Xiaohua, general manager of the same-side Internet applications industry, told China Electronic news Reporter: "government departments have data centers or information centres, and our first target is data centers or information centres." According to the depth of its mission consciousness, we divide the government's data centers into two types: one is aware of its commitment to the service-oriented government's mission, the need to use monitoring data and analysis results to support management and decision-making data centers, they are our first concern, the other is in this regard the gap between the data centres, We are mainly communicating with them on the idea. ”

The information resources of large data are very complex, how to extract useful information from these complicated information? Li Xiaohua revealed that the important thing is to set up a data resource system, the data into a hierarchy, including management data, decision-making level of data, while the data and the relevance of data annotation, not only with the form of data relevance needs to join , the correlation of different morphological data should be added to reflect the relevance of objective things. With such a data resource utilization framework, and then through with the management, business, social docking system, the formation of data resources system, so that data can really in the management, decision-making, monitoring, evaluation, human life to produce value.

(Responsible editor: The good of the Legacy)