Big data analytics technology is not meeting enterprise needs, according to 97 percent of data scientists polled by Revolution Analytics at the recent Joint Statistical Meeting.
Approximately 200 scientists participated in the survey, which identified the inherent complexities of big data software, problems applying valid statistical models to the data, and a general lack of insight into what the data means as three obstacles to running analytics on big data.
The survey offered no consensus on the definition of big data. Some data scientists said the big data threshold was a terabyte, others said a petabyte, and some said it was "just above what can be reasonably managed for any given job."
Technologies used to analyze big data include cloud computing platforms, massively parallel-processing databases, the Apache Hadoop Framework, and distributed databases. Gartner recently warned that big data is "heavily weighted toward current issues and can lead to short-sighted decisions that will hamper the enterprise's information architecture as IT leaders try to expand and change it to meet changing business needs."
From Network World
View Full Article
No entries found