Eliminate eight myths about large data

Gartner analyst Mark Beyer thinks: If corporate CIOs want to achieve large data normalization in the http://www.aliyun.com/zixun/aggregation/33722.html ">2020 year, Start by eliminating eight myths about big data.

Get 10 CIOs to define big data, and you'll get 10 different answers. This is because big data is still not standard for enterprise IT pros, said Gartner analyst Mark Beyer.

Beyer at Gartner's Symposium/itxpo meeting this year. "When things get very common, it starts to normalize, and our job as IT pros is to normalize big data 2020 years ago." ”

CIOs can help their businesses move toward normality by distinguishing facts from big data lies. "Myth helps to ease anxiety and is not conducive to reality," he said.

Here are the eight big data myths presented by Beyer:

1. Large data starts in terabytes. Do not look for large data standard size, because it does not have a standard size. "The big data is the processing of the data, not the size of the data," Beyer said.

2. If you want large data, you must replace the infrastructure. "If I decide to change the entire infrastructure because of new requirements, I'm betting everything I've had before," Beyer said. What are his lessons? "You have to figure out whether the risk of maturity (infrastructure) sacrifice is worth it." ”

3.80% of the data is unstructured. This may be the most frequently cited large data statistic, but according to Beyer, it is not accurate. "The world's largest information asset is machine data. Because they are not interconnected, it is absolutely a lie that they are unstructured. Machine data is structured data. By the way, these large numbers of machine data are often repetitive messages that confirm everything's normal. "That's what machine data usually says," he said.

4. Tools will replace data scientists. Rest assured that all the flowers are attracting, wooing, and getting data from the scientists ' money will not be white, Beyer said. "Tools are a kind of engineering, and engineering is the reuse of facts that have been discovered." And science is to discover new facts. "Tools do not replace data scientists-at least not until the tools can replicate and develop themselves."

5. More data will solve the problem of data quality. "The lower the quality of the data, the lower the quality of the answer," Beyer said. CIOs should be concerned about the quality of their data. For example, some people equate mobile phones with real individuals, he says. However, the phone can be left unattended in the office, or the GPS function can be turned off at any point in time. "Cell phones are not people," Beyer said.

6. Real time is just faster. Real-time operation does not mean speeding up the current data intake cleanup and analysis process, Beyer said. Instead, "Make sure that the shorter the interval between the data collection and the decision, the better," he said. In addition, most enterprise data does not require real-time operation.

7. The amount of data is superior to expertise. Those who think they can simply stop the business process, think again. That's because "a good data scientist has to be stopped at some point," Beyer said. Without business processes, data scientists will continue to continue without providing commercial value. Someone needs help to draw the line.

8. The data model is not available. This assertion is absolute. However, Beyer clarified that everything in any digital asset has a digital model. "We don't abandon models because of big data," he said.