Parsing: Where does the big data analysis platform start?

The term "big data" usually refers to data that is dramatically multiplied by quantity, speed and type. According to a new study by the Enterprise Strategy agency, the large data analysis platform is mimicking this definition: The number of vendor product releases is growing, product enhancements are rapidly doubling, and there are now multiple deployment options to support.

Julie Lockner, a senior analyst at ESG and author of the firm's Big data analysis platform, says companies are considering how they can integrate large data technologies into their architecture, especially when it becomes affordable and scalable.

Part of the misery stems from the fluidity of large data technology and terminology, which creates a tangle of market turmoil. Lockner her research as "Market Outlook Report 101", which she believes can be smoothed through internal evaluations and training.

Doing so means starting from scratch, starting with the definition.

Big data has a lot of choices.

According to ESG's report: "Big data analysis projects have sprung up, and some have not even understood the real meaning of big data to start doing." ”

Depending on the individual's understanding of this definition, the term's meaning is likely to widen or shrink. In fact, its definition has become very broad, and ESG has given its own explanation: "A DataSet that goes beyond the bounds and size of normal processing power forces you to take unconventional approaches." ”

The problem, says Lockner, is that the volume of data will grow to terabytes, and that "stress fractures" will begin to occur on the system, and that conventional techniques will not be able to guarantee cost-efficient methods in large data and large data analysis. That's when companies should consider expanding their data centers. Many big multinationals have been doing this before, but now there are more affordable choices. Whether it's a budget or a skill set. ”

At present, enterprises use a large number of large data deployment scenarios, there are custom development methods, large-scale parallel processing of databases, cloud computing services or a combination of some available tools. The addition of the open source Apache Hadoop project has sparked a growing interest in the open source project, which supports distributed processing of large datasets.

"I don't remember any other technology that can have such a big impact since the birth of HTML," Lockner said. ”

Vendors like IBM and EMC want to figure out how to integrate Hadoop into their product offerings. On January 9, for example, Oracle launched a large data machine that included a partnership with Cloudera, a Hadoop reseller. The situation now is that if a factory talks big data without mentioning Hadoop, you're embarrassed to post new products.

Although Lockner sees many of the promises of Hadoop and believes it will be in most enterprise data centers in the future, her research shows that it is still a new technology that should be used for specific scenarios.

Big Data started.

It is a good starting point for companies to explore investing in large data analysis platforms, to review the definition of large data by suppliers and to understand the relevance of their products to large data. Lockner said: "When you communicate with suppliers, to understand their product positioning and solve the problem?" ”

For example, EMC has a variety of large data products, such as Greenplum database software, greenplum data computing devices, and Isilon. These three products deal with different types of problems. "You have to really peel off the onion and do some homework," Lockner said. ”

First, Lockner recommends that customers rely on their well-connected suppliers to see a demo of their large data analysis platform. These are all free information. Because people in this business try to understand what they want to do, they should be able to put pressure on suppliers.

She recommends that clients also learn about case usage for other vendors in their industry. This information can help to see which vendors are real opinion leaders and which are not.

Companies should rely on their in-house IT departments and their more savvy employees to help with their homework. "In general, some lab projects will study new technology, and it's a pretty good start if companies can find those groups and brainstorm with them about how to do it," Lockner said. ”

But to really peel off these layers, companies should determine what the real demand is, and how the supplier's products can meet those needs. According to the report, this means estimating the internal available skills, where the data will come from, how quickly the behavior needs to be completed, and what needs to be integrated with the new platform. "Understanding business needs is more important than having excellent technology," says Lockner. ”

(Responsible editor: The good of the Legacy)