Too one star morning Feng Xiaojie: Analysis of four misunderstandings of large data

"Tenkine Server channel September 5 News" The current large data is very hot, but specific to what is large data, each manufacturer has different answers. We know that the industry relies on four characteristics to define large data: Volume, large volume, PB level; Produced, a wide range of data types; veracity, low value density; Velocity, fast processing. In short, you need to grab massive amounts of data from different dimensions and quickly turn them into ordered, available information.

At this stage, the main problem of large data solution is divided into 3 categories: expanding the traditional business intelligence (BI) domain. Previously for large data Statistics, association analysis, trend prediction from sampling to total analysis, the return of data to various reports, business process change. The aggregation analysis of various data is used to make business process improvement and evaluation basis, data commodity and commercial application. Data products or data services are formed by servicing or packaging existing data or processing capabilities.

Application delivery manufacturers too one star Morning product director Feng Xiaojie said that large data sheets from the literal meaning seems to be not difficult to understand, can be considered to be a massive level of data, but in this mass-level data on what it means, this in many industries and people in the concept of a pure in some misunderstanding.

Big Data misunderstanding one: As long as the big is good

Companies are facing massive growth in data volumes. For example, IDC's recent report predicts that global data volumes will expand by 50 times times by 2020. At present, the size of large data is still a changing indicator, the size of a single dataset ranging from dozens of TB to several PB.

A lot of people bring up big data, and if you don't mention the number of days processing data GB, Hadoop cluster has how many nodes, total storage How many Pb and other languages, are very afraid of others feel unprofessional. But, is really only the data big, is the big data?

Feng Xiaojie says, if the data is only big it is not very useful! Just as the meaning of money is how to use turnover, the data is big, but do not use, let it alone Pianan room corner, then it is not big data. For example, a lot of traditional portal sites, is basically sitting in the Golden Hill but no blessing of consumption situation. Hundreds of millions of users a day, but only a simple advertising presentation, not through the analysis of the data to produce more value.

Big Data misunderstanding two: only technology Daniel can understand big data

Large data can improve the processing speed of data by mapreduce this parallel processing technique. MapReduce's design is designed to achieve large data parallel processing through a large number of Low-cost servers, the data consistency requirements are not high, its outstanding advantage is extensibility and usability, especially suitable for the mass of structured, semi-structured and unstructured data mixed processing.

Traditional data management and business analysis tools and technologies are under pressure from large data, while new methods to help companies gain insights from large data analysis are emerging. These new methods are data processing, analysis and application in a way that is completely different from traditional tools and techniques. These new approaches include open source framework hadoop,nosql databases and large-scale parallel analysis databases (such as EMC's Greenplum, HP's Vertica). This means that businesses also need to rethink their approach to business analysis from a technical and cultural two perspective.

Feng Xiaojie says the application of large data is more of a strategic capability than a detailed execution skill that can help policymakers see the value of business opportunities from endless data and thus bring higher profits to businesses. And as decision-makers do not care too much in the technical details level, the big data exactly how technology generation, and how to straighten out the user experience.

Big Data misunderstanding three: It's a company that has big data.

is large data suitable for large enterprises only? For the general company, it is impossible to have PB-level data, and can not support the high cost of data storage, and large data technical personnel is very scarce. But small businesses can also take advantage of third-party data-processing service platforms. As to whether the enterprise needs large data to see its own business needs.

Feng Xiaojie said although the big data is a sweet pastry, but not everyone can digest, or not all have the necessary data, but to measure the status quo of the enterprise, see clear primary and secondary contradictions, or to consider good input-output rate of return, large data is not suitable for all the status quo of enterprises.

Big Data misunderstanding four: I want massive data

If there is a huge amount of data can help the enterprise development? This goes back to the analogy between big data values and money values. Obviously, the money that does not flow is more and more worthless, and the larger the base, the greater the loss that may result.

Money is so, so is big data. Only like Bitcoin players, keep using the data, and with unparalleled enthusiasm to dig the relationship and value behind the data, can be like a snowball, so that the relationship between the data richer and more perfect. Similarly, for the enterprise's large data, only the full use of large data, so that large data flow fully, and continuously realize the value-added effect, then there is a chance to release large data energy.

As a result, Feng Xiaojie points out, for business decision-makers, there must be a sober understanding of the big data, and before the head is ready to spend large sums of money, it has to be understood thoroughly: do I really need large data? Can large data really be harnessed for me?