How far is the operator from the big data?

In the wave of cloud computing, large data and mobile technology, IT systems play a more and more important role in telecom operators ' business, and transform from back-office support system to core business system. A survey from the Economist Information Division shows that more than half of the respondents rely heavily on it to improve their operational efficiency. Among all respondents, the most efficient companies, those that outperform their peers, have identified different roles for it in key areas of their business. 2013, operators will invest more money and human resources to transform IT systems.

In this year's It month report, we will focus on four aspects of big data, cloud computing, it centralization, and security, and how operators use these technologies to turn IT systems into a new engine for corporate performance growth.

Now it seems, "as if the spring breeze, thousands of trees pear blossom" to describe large data propaganda effect is not enough, and on behalf of its release value is a bit exaggerated.

Verizon started studying big Data last October. For Verizon, Big Data is a novelty and is in the exploratory phase. In fact, operators will not easily change their IT infrastructure until a technology has yet to produce economies of scale. Verizon consultant Yang Junjie said in an interview with the communications industry newspaper.

However, as a CRM consultant for Verizon's marketing department, Yang Junjie believes big data represents the future direction of the carrier's IT system. It was last October, when Verizon set up the Precision Marketing department precision Marketing Division.

At present, a number of consulting companies give analysis reports that the telecommunications and the Internet is the largest data as the two market segments. As a network operator, operators hold the most comprehensive information of users. The depth of use of these data will bring significant value to operators. A number of respondents told reporters.

"Anyone except God must use data to speak," said Edward Daming, an American management and statisticians. "Telecom operators around the world are using action to achieve the goal of" talking "online.

Pipelines are data mines.

The development of smart terminals has led to exponential growth in operator data traffic, with mobile Internet traffic growing by nearly 10 times-fold over the past 18 months, according to analysis reports. Another data report from Cisco shows that in 2001, the global IP stream to reach 1EB time is one year, and development to 2013, the time is significantly shortened only a day, 2016 again shortened to a half-day.

Operators ' networks are generating huge amounts of information every day. For example, Shanghai Telecom, Xujiahui business district generated 80,000 updates per minute information, the entire Shanghai city every hour has 3 million mobile phone calls, a daily 70tb-100tb of data access, 3 billion times a day, the amount of internet clicks. Obviously, behind these huge amounts of information is a data gold mine.

"All this information is generated in the operator's pipeline. Pipelines are the gold mines in which operators dig bokhary data. In this pipeline, operators will get three of the most valuable information, namely, mobile user location information, signaling information and network management and logging. "ZTE Big Data chief Engineer Ye in the communications Industry newspaper (NET) reporter interview said.

Location information has begun to benefit Telefónica. After setting up a large data unit, Telefónica launched its first product, "smart tournaments", which analyzes user traffic and key factors for a period and a location, and delivers insights to government and enterprise customers.

These insights can provide the retailer with decision support in the design and location of new stores and the way of merchandising, thus helping retailers to better insight into customer needs and improve marketing performance.

By modeling location and signaling data, operators can also use large data to support public affairs. He Hongling, a project manager of China Mobile, introduced that operators can establish user "time-space" feature analysis model, according to historical information and current information, to serve traffic planning, major events and emergencies.

Prior to the advent of large data, operators were more sources of data analysis based on BSS and CRM systems. "The data produced by the pipeline is more valuable than the BSS system," he said. "Ye said.

Teradata, general manager of China Communications industry, Wu Shunyu to reporters to further elaborate the importance of network-side data. Teradata, he said, worked with an operator to assess customer satisfaction with operator service. Based on the BSS data, the analysis results show that the user satisfaction is very high, the fee is timely, almost no complaints.

The reality, however, is that the operator's user-net rate continues to rise. Therefore, Teradata recommended the introduction of OSS-side signaling data for a comprehensive assessment. Signaling data show that many users actually often do not call, or frequent drop, download content speed is very slow, user perception is very poor, which is also the key reason for the user off the net.

The BSS system focuses more on marketing-type data. Therefore, operators should be large data, the BSS side and the OSS side data must be fused to form a full network of data. "Wu Shunyu said.

Shanxi Mobile then through the introduction of large network data, and BSS data fusion, accurate observation of micro-letter, mobile phone QQ, rice chat and other OTT business impact on the traditional business operators how much, and made the right market decisions.

However, to achieve the data fusion of BSS domain and OSS domain, operators are confronted with the challenge. Telecom operators ' It architectures are often likened to "chimney-style" architectures, where data islands are large. It is not easy to integrate these data into scale.

The chief scientist of the China Mobile Research Institute said on Weibo that "while telecommunications operators have a natural condition for analyzing user behavior, operators are based on the business development model of infrastructure (capability and technology) upward (market demand) and operators ' traditionally constrained by event-driven business architectures, It is difficult to produce direct market value so far. ”

In this respect, Wu Shunyu said, the big data solution that the manufacturer launches is making up this flaw.

Hadoop is not a universal key.

In the process of large data exploration, the operators are getting rid of the misunderstanding of Hadoop technology. While the big data is on the rise, Hadoop is seen as a panacea for the industry's challenge to big data.

However, in practice, it is found that Hadoop is inefficient in data mining and is far less useful than traditional data warehouse. But the low cost of Hadoop, as a data storage with the data warehouse unmatched advantages.

Jiang Xin, senior director of Teradata Solutions, uses a figurative metaphor to illustrate the role Hadoop plays in large data. "Big data is a process of gold panning. In the gold rush, the first need to dig out the ore, the need for transport tools and stacking positions, then Hadoop is equivalent to this tool and location, more data transmission and storage. After that, the ore needs to be refined to make gold. With Hadoop, the cost of human and material will be greatly increased, and the time costs will increase. ”

The large data application representative of China Unicom, "the support system for the query and analysis of network records of mobile communication users" is the function of using Hadoop technology to realize real-time query. Can handle 70 billion online records per day, the user online query records of the system response time shortened to a second, greatly enhance user perception.

From the current situation, any single technology is not enough to support the large data structure efficiently and cheaply. In a speech at the "2013 Teradata Big Data Summit", He Hongling pointed out that in the big data age, operators needed a rigorous mix-and-match structure, including traditional data warehouses, new data repositories, and Hadoop to give full play to the advantages of each platform.

Moreover, he stressed, these technologies to effectively form a "mix", rather than "disorderly", can not become a "shaft", need to integrate into a unified large data architecture. At present, China Mobile has formed a unified large data structure based on traditional data warehouse.

Among them, the main data warehouse is used to store "hot" data, perform efficient data processing, support KPI, timely presentation of monthly report, high reliability and low delay access to support data, and the new relational database as the depth analysis platform, the user stores "temperature" data, supports all kinds of advanced analysis and data mining; Hadoop is used to store "cold" and unstructured raw data, to achieve custom, flexible data processing and mining, and large-scale simple data query work.

Two major profit directions

Large data as the carrier to be excavated gold mine, its value comes from two aspects, one is to improve the user experience, for users to achieve precision marketing, improve operating profit; the second is to provide the results of data analysis as a service to enterprise customers, to help them achieve more value.

Ye told reporters that DAAS (Data as A Service) to bring more value, more worthy of operator attention. As an important data aggregator in the large data age, telecom operators can open up data to form large data value ecological circle, and fully tap its potential value. "he said. The Spanish telecom's wisdom footprint product mentioned above represents the second direction.

It is worth mentioning that operators need to solve the problem of user privacy. "In recent years, the phenomenon of customer privacy abuse has occurred. In the big data age, this threat is being magnified. The protection of customer privacy is not only the compliance of laws and regulations, but also the social responsibility of enterprises. "He Hongling said.

Verizon has adjusted its privacy policy in the year before it set up a large data unit, laying the groundwork for legitimate use of user data. In a statement on changes in privacy policy in October 2011, Verizon said the company would "write business and marketing reports" and "create mobile ads that are more relevant to users", using information they have collected from users ' websites, the applications they use, and their geographic location. Verizon is also prepared to share these with other companies.

But Verizon says users who are unwilling to share their information can opt out. If the user decides to participate, the information it uses or shares will not be recognized as a user's information.

The study, published by IBM and Oxford University, "Analysis: Big data in the real world" shows that 25% of Chinese companies are already doing big data practices, and 56% Chinese companies are planning to deploy big data. It is expected that in future years, its value will be released in size.