Data explosion-type growth Internet introduction of large data technology is imminent

Data produced by Machine-to-machine in 2005 accounted for 11% of the world's total data, according to IDC, which is expected to increase to 42% by 2020. This proportion shows that the rapid development of the Internet of Things has also produced a large number of data, how to properly handle and reasonably use these massive data is the key to the next development of the Internet of things. "In the recent 2013 (fourth session) China Internet of Things meeting held by China Electronic Society, academician Hequan of Chinese Academy of Engineering strongly appealed."

However, whether the large data, or the Internet of things in China are in the initial stage, as far as the intersection of the two areas people are involved in very little, and the complexity of the data generated by the Internet of things, it is further increased the field of the Internet of things large data landing difficulties.

Geometric growth

Figuratively speaking, the internet of Things is the process in which objects perceive and react to the outside world, and in this process, both the perception of the outside world and the reaction are related to the generation and processing of data. Especially in some areas, such as security, mobile medical, intelligent transportation, but also because of the application of the particularity will produce massive data.

Take security as an example, in order to effectively detect anomalies, the system needs to monitor the region for a 7x24-hour monitoring, which will produce a large number of video data, for hindsight, these data often need to save a period of time, in addition, the system also needs to compare the data at different times, To be able to react in a timely manner when an anomaly occurs. In the field of mobile medicine, in order to achieve the sharing of medical resources between different places, the standard of diagnosis of disease is indispensable, so many hospitals have advocated for CT examination of patients, and a patient's CT will often reach dozens of GB, according to the annual number of inpatients in China as many as 200 million people, and patient data usually needs to be retained for more than 50 years, and the amount of data produced by mobile medicine is considerable.

What makes the situation worse is that, in the case of the ITU, the data needed to be analyzed by the Internet of Things is not only the data collected by the sensors, but also the data generated by the virtual objects, because only by placing the Internet data in the background of the data generated by the virtual network can the larger data value be mined. The data comes mainly from social networks and government websites. Hequan said that the Teitter site added 200 million micro-blogs a day, Facebook uploaded 250 million photos a day, by 2015, about 1 people to spend 5 years to see the internet for a second of all the video, this shows that the amount of data on the internet is huge.

Despite the massive growth of the data, the processing power of the data has not increased synchronously. "The world's new data is growing by 40% a year, and global information can double in two years, with less than 5% per cent of the processing power and utilization of the added data." "Hequan said. Therefore, speeding up the introduction of large data technology to promote the development of Internet of things is imminent.

Large data storage particularity of IoT

IoT operators also recognize the importance of large data applications. "Based on practical experience, it is true that in many applications, the Internet of Things is not only a single thing networking technology, but also an effective integration of large data and cloud computing technologies," said Li Anmin, dean of the Shanghai Institute of China Telecom. The popularization of IoT will bring the city into the era of large data, we need to pay attention to the application of large data needs, from the technology, business, business models, etc. in advance to deal with. ”

However, the application of large data in the Internet of things is not so simple. This is primarily determined by the characteristics of the internet of things.

Hequan that the large data generated by the Internet of things has different characteristics from the general large data. The data of the Internet of Things is heterogeneous, diverse and noisy, and its growth rate is very high, which increases the difficulty of data processing analysis. In addition, the data of the Internet of things has obvious granularity, its data usually has the time, the position, the environment and the behavior and so on information, the Internet of Things is also the social data, but this is not the communication information of the person and person, but the social cooperation information of the thing and the person. Therefore, the blending of IoT makes it easier for the system to make more scientific decisions by combining the data of the Internet with the data obtained through social media and the information released by the government.

There are some obstacles to the development of large data industry in the field of Internet. Although the concept of large data is hot in recent years, however, the development of large data shows imbalance, on the one hand, the rapid development of technology, on the other hand, how to obtain value from large data, is still in the discussion stage. In particular, telecom operators, although also recognized the price of large data, but for their own in the large data industry positioning, as well as large data business model, now has not formed a clear and unified understanding. This situation restricts the development of large data, which also affects the application of large data in the field of IoT.

Large data and Internet of things can drive each other

Considering the urgency and relative backwardness of large data in the field of IoT application, Hequan that it is urgent to speed up the introduction of large data. For the characteristics of massive data, Hequan that although data resources are very important, many of the data is complex and redundant information, such as in video surveillance, people need to focus on the video information in the abnormal events, rather than the regular, unchanged data, so take into account the analysis of historical data and new data, Compress the duplicated data and focus on the burst data that analyzes the exception.

Granular and unstructured data is a major feature of the large data in the Internet of things. Although the data processing of this kind of characteristic is very complex, but they are crucial to make the correct decision of the system, so parsing unstructured information is an important link that can not be neglected. In addition, unstructured data can be structured to unify the data of different architectures across different systems, through the Unified IoT architecture design.

The sharing of data information among different departments contributes to the greater value of the Internet of things, so hequan that, while it seems unlikely that different industries such as electricity, transportation and industry can be integrated into one thing, it is feasible to share data from different industries.

And for the relatively slow pace of large-scale data commercialization, some experts believe that on the one hand, the development of Internet of things requires the industry to speed up the study of large data application models and business models, on the other hand, the application of the Internet also helps to clear large data business model, for large data development opportunities.