Industry perspective: Large data in Hortonworks eyes

The hazy definition of the current spread is not enough to articulate the benefits of big data, says an executive at the Hortonworks company. Today we are going to look at what the big figures in their eyes are all about, from the perspective of the industry.

So what is the big data? This general technician uses the classic 3V model to explain--capacity, speed, and data diversity--that is almost an industry convention. But the popular definition is too vague to really explain the real benefits that big data platforms bring to users.

David McJannet, vice president of marketing at Hortonworks, argues that a more realistic description of the benefits of big data to the real world is more conducive to extending the new mechanism to all walks of life.

"Big data is by no means something that is not clear," mcjannet in an interview with reporters. "From a pragmatic standpoint, this is a new type of data that companies have not previously focused on, primarily as a basis for the operation of new analytical applications." ”

Of course, Hortonworks's move to promote a clear concept of big data around the world also has its own considerations. As the main catalyst in the Hadoop ecosystem, the california-based enterprise software company is able to convince business users to store and analyze large amounts of data to help them sell their products and make a profit, and this emerging sector has been overlooked by customers in the past.

So they propose an alternative definition (from an objective point of view): The purpose of large data is "to build new analytical applications based on new data types, to better serve customers and promote competitive advantage".

This seemingly simple definition can help companies "go beyond previous fuzzy understanding of big Data".

Of course, there is no similarity between the big data, so hortonworks companies classify five different data categories based on their specific sources: social media, server logs, Web click streams, devices/sensors, and geography.

But how do enterprise users use this information?

Look at social media data first. Companies are now using Facebook, Twitter, and such social networking sites to learn about the "mood" users have about something, McJannet told reporters. For example, a filmmaker can learn about the evaluation of new films based on such data and optimize the marketing campaign based on the views of social media users.

Server logs Help system administrators use Hadoop to discover data to identify and address critical issues. McJannet The example: "If I track every single inbound request on my site and overlay it according to the geographical zoning, I can better judge where my large customers are concentrated and where they may face potential security issues." ”

The click Stream data brought by Hadoop can help users manage the overload state information of traditional data management system efficiently.

"If I can capture all the streaming data from my website--and, of course, such huge data records will quickly fill up the existing database--the data generated by sheer clicks," McJannet explains, "then keep it in Hadoop ... will help me create a very interesting profiling application based on the information. ”

Device data is also a large part of the untapped data source.

"The device is definitely one of the largest sources of data, covering a wide range of common areas such as air-conditioning units, refrigerators, lorries and even household machinery," mcjannet pointed out. "Such processes will lead to explosive data growth. ”

At present, the world's mobile phone to reach billions of, so mobile data acquisition equipment has a broad market development space. "Every time you go through a call, the information between the towers is converted, and some pieces of data are generated. If someone is going to create a profiling application, that information can be a valuable material base, "McJannet said.

Geographic data is also not as long, until 10 years ago only in space technology and military applications. Now it has identified a new way of development for commercial applications.

For example, the transportation company can track the location data of each vehicle every 10-60 seconds, and accumulate PB-level related information.

"If you plan to use geographically relevant data in your business processes, you should first consider what applications you can create and what valuable information you will be able to extract from them," McJannet concludes.