Big data has changed from technology-driven to application-driven.

Big Data: From "Technology Driven" to "Application Driven"

After the Internet of Things and cloud computing, big data has become one of the most concerned concepts in the current information technology industry. With the advent of the era of big data, the boundaries between fields and industries have become increasingly blurred, application innovation has surpassed technology itself, and production methods have changed to services. As an asset, data brings new business value to enterprises, and data openness brings opportunities and challenges to government governance and personal well-being ... No matter individuals, business organizations, social groups, or countries and economies, the dream of big data can be realized.

At present, the global big data industry is in the gestation period and opportunity period of vigorous development. The development and upgrading of core key technologies have been accelerated, and various solution providers have stepped up their publicity, especially drawing beautiful blueprints around key areas such as telecommunications, aviation, transportation, biology and urban management, making efforts to promote industry applications and business model innovation and seize industrial growth points. At the same time, small and micro enterprises and entrepreneurs are passionate about big data and hope to take this opportunity to realize their dream of rapid growth. As the whole big data industry begins to turn to the stage of application innovation, the expectation of high growth makes all parties optimistic about the future.

From "technology-driven" to "application-driven"

As an independent industry, the industrial system framework of big data presents the characteristics of "two verticals and three horizontals": "two verticals" are based on the basic technology level and are divided into the underlying technology and the application technology. The former is * * * and basic technologies, such as Hadoop framework, Hbase database and Mahout algorithm set. The latter is a kind of "secondary development" behavior, including all kinds of personalized programs, products and services. Based on the process sequence of processing, "three horizons" can be divided into infrastructure, analysis system and application tools, and can also be further refined into five aspects: data collection, storage, processing, analysis and service. At present, the industrial system of "two verticals and three horizontals" is mature, which can meet the needs of most industrial applications.

The generalized application of big data is essentially a kind of "value-added analysis", and its prospect is almost unlimited, and it is not limited by any industry, resource, region or user. From this perspective, almost all the future development directions of various industrial sectors can be linked with big data. Taking the "Twelfth Five-Year" national strategic emerging industry development plan as an example, the description and layout of many technological frontiers are consistent with or related to big data, or can be realized through big data. For example, the new generation of information technology industry has laid out the Internet of Things, mobile terminal equipment, cloud computing and massive data processing software; The energy-saving and environmental protection industry has laid out efficient energy storage, energy-saving monitoring and energy metering; The biomedical industry has laid out biological resource sample bank, gene sequencing, and remote health management services based on the Internet of Things.

Since the emergence of big data technology in the Internet era, the rapid development of the Internet is closely related to its concept of openness, sharing and cooperation, so the innovation of big data technology has also introduced this value of the Internet. For example, many big data technologies are open source and can be used and improved by developers all over the world for free. The maturity of open source projects, open source communities and open innovation alliance organizations has promoted the development of core technologies of big data and spawned various new products for storing, processing and analyzing big data. This process effectively reduces industrial technical barriers, promotes the participation of more enterprises and entrepreneurs, further accelerates the process of technology application transformation, and contributes to the rapid growth of the industry.

Although the "technology-driven" color of the big data industry is very obvious, there is still a distance from the "application-driven" stage, but this transformation process is accelerating.

Detailed industrial competition strategy is gradually taking shape.

Big data industry is a typical knowledge-intensive service industry. Except the infrastructure link will bring a certain amount of energy consumption, other links are zero energy consumption and high added value. Its entry threshold in initial capital and supervision is extremely low, but it requires higher human resources. Therefore, industrial competition presents the characteristics of large quantity and high level, and the competitive strategies of enterprises are gradually divided.

Although the number of big data practitioners is rapidly increasing-almost all information technology companies are deployed in this field, entrepreneurs are constantly entering, and there are many competitors, but this brings not excessive competition, but healthy competition, which will eventually promote technological innovation and value realization.

This is mainly due to two reasons: First, the high innovative nature. Big data technology is a high value-added link in the information technology field. Big data companies represented by Google and Amazon are leading the world in technological progress, innovation activities and market share. The second is high growth expectations. As individual enterprises, under the expectation of rapid industrial growth, they basically choose the strategy of pursuing specialty and winning by product performance and service, and give up the strategy of pursuing low cost.

In the process of competition, different types of competitors have their own advantages. According to the innovation and application level of technology, it is mainly divided into three types of competitors: one is "Internet subversive". Google and various big data open source projects have developed brand-new basic technologies and database architectures. Relying on the so-called free and open source Internet model, it completely changed the original technical standards and rules of the game and subverted the fragmented information technology industry in the past.

The second is the "newborn calf". Facing the new regulations, big companies and entrepreneurs stand on the same starting line. Some entrepreneurial enterprises with core talents and market awareness quickly seize opportunities in specific tools and professional platforms, fill market gaps, achieve rapid development, and occupy a place in the industrial chain.

The third is "system integrator". Traditional IT giants such as Microsoft and IBM have abundant capital, R&D capabilities and market resources. They can be keenly aware of the urgency of self-revolution and immediately take countermeasures, actively acquire big data-related enterprises, assemble the acquired technical products into industry-oriented application solutions, and strengthen the commercial marketing of big data.

In addition, the government is also an important part of the big data industry, which is mainly reflected in the openness of the government to public data, which will enable the government to play a more important role in promoting industrial development.

In 2009, the first presidential memorandum signed by the newly elected US President Barack barack obama was transparent and open government, and then Data.Gov, a unified open portal for government data, was established to gradually open public data owned by the government, and various application interfaces were provided for developers to create featured applications. By the beginning of 20 14, the website had opened more than 85,000 data sets and collected more than 200 applications, software tools and mobile phone plug-ins of/kloc-0, of which more than 300 were developed by individuals or non-governmental organizations. New business models and enterprises have emerged. For example, FlightCaster Company provides flight delay forecast based on the data of US Transportation Statistics Bureau, FAA Traffic Control Center Alert, US Meteorological Bureau and flight operation information website FlightStats, which is 6 hours ahead of the official notice of airlines, with an accuracy rate of 85%-90%.

Data-driven business model innovation

Data-driven business models have sprung up all over the world. According to the distinction of data acquisition, management, analysis and application, the business model of big data can be divided into three types: data hosting and trading platform, relationship mining and precipitation value utilization, data socialization and cross-border connection.

The data hosting and trading platform model has been applied for decades and is the most mature and universal big data business model. Its essence is to play the scale effect and reduce the investment cost of a single enterprise in data information storage and search. The main business forms are space leasing and hosting, data store, data market and so on. Typical representative enterprises are Amazon, EMC2 and DropBox.

In recent years, the concept of "cloud" has been introduced, from simple data storage to data aggregation platform, and finally cloud service has been formed; The integration of unique data resources is developing towards vertical industry consolidation and horizontal multi-industry integration, which promotes the emergence of one-stop data stores and data trading platforms. Companies such as Amazon and Microsoft have established data stores that can trade applications and advanced data sets. At present, there are trillions of data points, thousands of subscriptions and hundreds of applications.

Relationship mining is the mainstream big data business model and the main application model of data science. The core is to find hidden correlation through data, which is finally used to guide business, accurate service and assist decision-making.

To realize this model, there are some preconditions, mainly data-oriented processing and analysis: first, complete quantification of target areas, such as Internet advertising, with complete and detailed data records from advertising clicks to user purchase behavior; Second, the data processing ability has been greatly improved, and it needs to be able to handle non-relational data and maintain real-time and fast performance under massive conditions. The difficulty of this model lies in the need to subvert the conventional user thinking and demand logic. The typical type is precipitation value utilization, which uses some usually meaningless data or even junk data and finally draws valuable conclusions.

For example, Google uses the misspelling records of billions of users to improve the intelligence of its spell checker. At present, the big data model based on relationship mining is not mature, but it carries high expectations from all walks of life: this model will help to drive industrial transformation and develop emerging industries, such as promoting the transformation of R&D-intensive industries such as biomedicine and knowledge-intensive industries such as enterprise consulting to data-intensive industries, promoting the transformation of traditional service industries such as retail and transportation to modern service industries, and promoting the transformation of traditional manufacturing industries to intelligent manufacturing industries.

Different from the first two modes, the data socialization and cross-border connection mode directly face every social individual, which is essentially to fully tap the individual resources of the physical world, turn it into a node of the virtual world, and connect, interact and trade with other nodes, thus greatly reducing the promotion cost of various commercial services and forming a new format. This model is maturing, and the most typical one is O2O.

For example, WeChat has become an important entrance to connect online and offline and carry out mobile payment; Taxi software effectively reduces the information asymmetry between the supply and demand sides and improves the intelligence of the taxi market; Wearable devices further quantify human body information and provide decision-making suggestions; Apple Passbook software provides users with a smart electronic card package. There are several necessary conditions to develop this mode, mainly data acquisition and transmission: mobility requires intelligent terminals to have location service and can transmit wireless signals; A stable connection requires a high-speed, ubiquitous external network environment; Online payment, relying on the user's final payment behavior to achieve profitability; Continuous sensing needs advanced sensor technology, low-power chip technology and battery technology as a guarantee.