IBM Puredata push Large data machine for Hadoop

With the development of large data projects in enterprises, the importance of data analysis speed is becoming more and more obvious. To further enhance the speed of large data analysis, IBM launched a large Hadoop data machine in California in Wednesday, designed to help enterprise users realize the need for more kinds and larger data (lower cost) real-time analysis.

Bob Picciano, general manager of IBM's software information management department, said:

Companies are being besieged by large data floods, and as manufacturers IBM must provide customers with better tools to dig bokhary data, which must be fast enough to handle massive amounts of data and be easier to use.

In Gartner's March bi Magic Quadrant, IBM is the most visionary and innovative "visionary" data analysis technology provider for all vendors, and IBM has two main ways to increase the speed of large data analysis of the enterprise (for more types). One is through the analysis of accelerated technology to transform large data into "medium data" or even "small data" to reduce workload. READ: Forget big data, start with "medium data", another way is hardware optimization, the introduction of Hadoop-oriented large data machines.

To achieve this goal, IBM released the Blu Acceleration Analysis acceleration technology. When the IBM user (for example DB2 user) runs the query, the blu can quickly narrow down the data analysis scope, after cleaning the massive data, only a small amount of valid data enters the analysis process, thus the user's data analysis load will be greatly reduced, the hardware processing resources can give the analysis result more quickly. An important function of Blu is data skipping, "skipping" large amounts of unwanted data and singling out important "small data". In addition, Blu can keep data compressed throughout the analysis process, which will greatly improve the data processing speed. IBM claims that in some tests, the addition of Blu made the analysis of the DB2 database 1000 times times faster.

In the second path of large-speed data analysis-hardware optimization, IBM has also released IBM Puredata box designed for the Hadoop data-processing platform.

As the third member of the IBM Puresystems family, the Puredata BAE large data specialist integration System was positioned by IBM as an analytical processing engine for the large data age, and last year IBM released the IBM Puredata system for transactions, analysis systems IBM Puredata System for analytics and Operations Analysis systems IBM Puredata System for operational Analytics (based on Netezza) three large data-machine products.

Puredata system Big Data machine products for Hadoop will be available later this year, according to Phil Francisco, vice president of product management and product marketing at IBM Netezza. The user will be able to load the data in 90 minutes, before the enterprise takes two or three weeks to complete data loading in the data center Hadoop instance. IBM's large data analysis function for Hadoop stores data that Hadoop has processed locally, which can meet the requirements of some industry customers for data jurisdiction.

(Responsible editor: Fumingli)