2017 Big Data market total will exceed 47 billion dollars

Although large data-related technologies, such as Hadoop, NoSQL databases, and memory analysis, are new to many people, it has to be acknowledged that these technologies have been used and developed more widely over the past year or two. How big is the big data? Jeff Kelly, an analyst at the Market Research institute Wikibon, said the 2012 Big data market was 11.4 billion trillion dollars and is expected to grow to $47 billion by 2017.

Jeff Kelly worked for TechTarget and worked as a news editor for many years. Recently, he has been interviewed by reporters, in response to the wikibon of large data forecasts for in-depth exchanges. This article is the interview record, mainly discusses the big data spread, the big data ecosystem and the NoSQL and so on topic.

When you study the big data market, what other elements are within your scope besides Hadoop?

Jeff Kelly: We've been thinking about this for a long time. Hadoop is certainly one of the most important, but not all. The big data is almost every aspect of data lifecycle management, so we finally decide to consider the new technology. Because traditional data management technology can not deal with massive, many types and high-speed flow of data, and Hadoop, NoSQL and so on.

We also evaluated the hardware because it was the basis for supporting the technology. In addition to some not very new, but with large data load close to the software technology is also the scope of our consideration. For example, you need data integration software to import data into Hadoop. But there is not a unified large data platform standard.

Traditional business intelligence reporting software is not the scope of our consideration, but some other visual software we will focus on. Big data is not just a technology, but an attitude toward data. With large data, people are starting to use some of the original tools to explore more ways to process data.

What are the big data use cases we see?

Kelly: As you can see, internet companies have been at the front of big data applications. Google has created the big data we know today, and Yahoo has created Hadoop, and other Internet companies are making new attempts. But we should also see that large data applications are being carried out in many traditional industries, such as the biopharmaceutical sector, which uses large data for genetic research.

Financial services companies are also early adopters of big data, such as leveraging Hadoop technology to gain more competitiveness. In addition, in the retail industry, although the pace of development is slightly slow, but they are also constantly trying to make data. Although people may think this is a bit exaggerated, I have to say that the big data has penetrated all walks of life.

According to your point of view, what is the big data market for large manufacturers such as IBM and Oracle?

Kelly: There is no doubt that big manufacturers are investing in the development of large data products because they know it's what customers need. Many manufacturers have started to force a large data market, and IBM in which has been a long time to go. IBM bought a lot of analytics software companies long ago, and they're still focused on so-called smart-earth projects, where big data or how to make better use of data is their focus. IBM is also a leader in the revenue side of large data businesses, and the blue giant's revenue for big data applications in 2012 has surpassed $1 billion trillion, according to Wikibon data.

Oracle uses Cloudera Hadoop in their integration systems (Reference Links: Oracle and Cloudera executives talk about large data machines), they also have their own NoSQL databases, and a wide range of Exadata All-in-one machines. There have been heated discussions about whether these can be big data strategies. Depending on the type of load you see, we'll call it big data.

But Oracle should be the most awkward of all vendors because the "hadoop/nosql" model needs to be scaled horizontally, and clusters should be made up of inexpensive PC servers. Oracle offers "big boxes" that are vertically expanded and expensive. How Oracle responds to this is highly noteworthy. Oracle's leadership in the database sector is ingrained, and this is their biggest asset, but we should also see that their current model runs counter to the open source community, which is the main driver of big data development.

What is the development trend of NoSQL, Newsql and Hadoop?

Kelly: The biggest trend is that Hadoop vendors are now trying to bring SQL functionality into Hadoop and NoSQL. People know that NoSQL is good at scaling and processing unstructured data, but to get it up to the enterprise application level, you must meet enterprise-level availability and security requirements. Now the trend of NoSQL is to transfer to the enterprise level, which will make the development of NoSQL database technology a new step.

(editor: Fumingli)