Big Data 2013 trend: SQL Reverse attack

The big data is "soar" in 2012, and it will change every aspect of data management in a dramatic way. Large data systems have brought changes to machine-generated data management, continuous ETL, operational bi, Dynamic Data, and cloud-based data warehouses.

However, with big data coming into the 2013, there is no technology that is more active than the NoSQL database and Hadoop, and they all have greater room for improvement. According to a 2012 report by Marketanalysis.com, the Hadoop mapreduce market alone is expected to achieve a compound annual growth rate of 58%, which will reach $2.2 billion trillion in 2018.

NoSQL and Hadoop appear primarily for unstructured data, such as text data or web logs. Like Apache Hadoop, these technologies are often launched from open source and then gradually become new commercial products.

Judith Hurwitz, President and CEO of Hurwitz and Associates, believes that large data architectures and large-scale parallel processing greatly alter the status of data applications. "Before that, even if the data really mattered to the company, people weren't really capable of acquiring massive amounts of data and analyzing them in real time," she said. Now, the goals that could not be achieved are within reach. ”

The attack of the SQL camp

We can see from the TechTarget Business Intelligence website that from 2012 onwards, the comments on the mainstream relational database will decline. Observers believe that the SQL relational database will lose its competitive edge in dealing with large data over the next few years, as compared to the techniques mentioned above.

The driving force behind this trend is that companies want to get more unstructured data at a faster rate so they can rely more on data-driven decision-making. Accustomed data processing methods are undergoing change to better integrate new technologies.

For the traditional relational database vendors, there are many examples of embracing big data and Hadoop over the past year:

IBM has been acquiring a number of advanced data analysis companies to augment the large data product line. The blue giant's efforts have been improved from small (for example, NoSQL graphics storage for DB2 10 and Infosphere Warehouse 10) to subversive Puredata All-in-one to help customers get big data.

Oracle launched a large data appliance in early 2012, and recently released Oracle NoSQL database version 2.0, which has been automatically rebalanced, the new application programming interface can handle large objects and be more tightly integrated with Oracle databases, and can also support direct SQL query Oracle NoSQL database records.

Microsoft showed a preview of Hadoop support for Windows Azure and Windows Server, Teradata Company released its aster large data analysis products, and Informatica released a large data version of the PowerCenter suite. It is said to eliminate the need for the manual coding of Hadoop, which brings programming tasks into the Informatica development environment.

Big data will never have a problem with who replaces it, although SQL has been hit in the past year, but it is not going to decline. On the contrary, some of the more professional companies in NoSQL and Hadoop have done a lot of work on SQL. A typical example is the Hadoop startup Cloudera, which promotes the degree of collaboration between Hadoop and SQL through Impala. Impala is a Hadoop software product that supports standard SQL for interactive queries.

Large Data Transformation

Big data changes are also driving the database technology forward, and now we see that SQL and NoSQL from the confrontation more to the fusion. In fact, in the early discussion of large data, traditional relational database technology is neglected.

"In the past few years, with the rise of big data, SQL has suffered a certain amount of impact as a mainstream technology," said Ronnie Beggs, vice president of streaming media database manufacturer Sqlstream. Larger data are associated more with NoSQL. ”

He's supposed to see a noticeable change in 2013. In recent years there have been a lot of efforts to make the NoSQL database better adapted to the SQL database style.

Beggs said: "The big data is constantly changing, and what we see next year is the return of SQL, which will serve as the interface for all large data platforms." ”

The coexistence of Hadoop, NoSQL, and SQL represents a new step towards maturity for large data. With the opening of the 2013, large data may have shifted from a hot topic to landing practice.

"I think people are trying to really understand their business value through the hype of big data," says Colin White, president and founder of the Ashland BI Research Institute. In 2013, I think we'll see better examples of people getting business value from big data. It's not about big data, it's about what you do with big data. ”

While there is a wide range of interest in new technologies, the speed with which different companies accept comprehensive data systems varies.

One system integrator told TechTarget reporters in the financial industry that banks, as a sector, had only partially dabbled in basic, large data, not all. Banks and other sectors saw only large numbers of data, without noticing its non-structural nature. At least for now.

"There are two parts to the meaning of big data," he said. The first part is that they are very large, and the second part is that the data is unstructured. Banks are clearly part of the first section. But we are not going to collect social data such as tweets, at least not yet. We're still watching to see how other users in the financial data services market handle it. ”