Big data players ' predictions of 2013 trends

This article comes from ZDNet reporter Adrew Brust's blog post. This paper mainly states the forecast of the development of the 2013 and some viewpoints of the author by the large data industry.

In large data areas, while forecasts for the technology industry are indispensable, adrew is obviously uninterested in promoting a specific company, and many industry companies have sent their plans to adrew next year, so he thinks: if you can integrate these 2013-year forecasts into your own perspective, Would be a very interesting thing.

▲ Figure: What stage will big data develop in 2013?

2013 of Hadoop

Start with Hadoop. As a representative technology for large data areas, many companies focus next year's plans on Hadoop. MAPR's executive director, John Schroeder, predicts that "user optimization of Hadoop will focus more on hardware". At the same time, pervasive Software's technical director Mike Hoskins also pointed out that "the enterprise-friendly Hadoop technology market will reach an unprecedented peak."

Adrew that both predictions would happen, and that, as a whole, Hadoop itself would have developed rapidly, and that Hadoop's configuration and seamless integration in multiple data centers would also be popular. Perhaps because of this, MAPR's Schroeder will say, "Hadoop's expertise is growing fast, but the talent is still scarce", and he predicts that "sql-based Hadoop tools will continue to evolve". There is no objection to the latter prediction, because in the past quarter alone, such a tool has grown enormously and the current momentum has waned.

However, not all viewpoints are consistent. Rainstor has a hybrid based on Sql-hadoop, predicting that once companies need to manage big data, they will look for new technology next year, not just Hadoop, which contrasts with pervasive's predictions. MAPR will occupy the traditional data Warehouse market, Hadoop will become a big data analysis of the choice.

Hadoop Upper

Hadoop has brought us a lot of subtle predictions in the past. "The big data enthusiasm of companies will be released from the underlying research, and they will focus more on a clear range of business and technology," Rainstor predicts. "MapR predicts," open source (increase revenue) application will surpass throttling application. Other companies ' forecasts basically assume that next year's users will move to a more sophisticated, larger data model.

Although they are right, it is not easy to move from the bottom, and many problems remain unresolved. Many forecasters seem to feel the same way, and the general argument is that Data quality will continue to be a hot potato for businesses. Progress DataDirect's experts predict that "businesses will be overwhelmed by their own data, and that scattered data will create new problems." "These predictions seem a bit alarmist, but in a world of business intelligence, these predictions are no longer new." If the big data world can avoid these problems, it will only be a good vision under the hype. Once these hype is over, those problems will be highlighted, so you can't escape.

Data analysis and Virtualization

Rainstor, pervasive, mapr and Progress DataDirect are not the only predictors of adrew predictions by industry insiders. Adrew also received forecasts from BlueKai and iOS data virtual supplier Roambi, which focuses on market analysis. Their predictions are based on their services, which generally involve the user's topic, pointing out that they are bound to encounter more and more complex problems.

For example, Roambi said, "Half of the business data is not available, which affects the accuracy of decisions and forecasts." BlueKai's executive director, Omar Tawakol, points out that "visionary brands reassess their proxies and measure data utilization and ROI based on maximizing the data capabilities of the cross channel audience." ”

Reduce complexity

The landing of large data technology will have two features: one is less dependent on mapreduce, the other is the application of Hadoop technology into the enterprise software architecture.

For the first feature, adrew refers to software such as Cloudera's Impala and Microsoft's polybase that are fully developed, bypassing mapreduce and processing data directly from the HDFs. MAPR predicts that the sql-based tool will continue to evolve with the software above, pervasive predicts, "yarn is changing the rules of Hadoop, which not only allow MapReduce applications to run on Hadoop, but also allow other types of applications to be deployed on top of it." ”

For the second feature, the large-scale use of Hadoop is an inevitable trend, and gradually will form the industry standards, and then become a more valuable software base, not only for their own internal use. So, Adrew thinks we'll see more business intelligence and intelligent analytics tools that integrate the Hadoop feature, so our reliance on the Hadoop experts will gradually come down.

See

Let's wait and see which of these predictions will come true.