Cloud computing era: when big data experiences agility

There were two major voices for big data technology at the o'reilly Media Conference in New York in September this year: enterprise level and agility. We know that enterprise-level business intelligence products include Oracle Hyperion, SAP businessobjects, and IBM cogonos, while agile products include qlikview, tableau, and tibco spotfire.

If it turns out that big data must purchase enterprise-level products, it means that big data will spend a lot of money. However, this is not absolute. By using agile big data technology, enterprises of all sizes can control costs and benefit from big data. It is vital to minimize costs and maximize understanding of big datasets. Once the data is converted to usable, it will have insights into the business and then summarize the problems in various ways, and give full play to the advantages of enterprise technology to solve problems.

Enterprise
Agile bi

First, let's take a look at what happened in the Bi world. Enterprise-level Bi is designed to meet the requirements of many information silos in large enterprises. A typical scenario is as follows: Data in the information islands is cleansed and normalized through the ETL process and enters the data warehouse to become an available version. Then, data is sliced, sliced, and delivered to thousands of people through report and analysis technologies. Enterprise-level Bi is a complex process. It works collaboratively with multiple applications to meet the needs of thousands of people in the enterprise. The problem with enterprise-level Bi is that it takes a long time to configure and modify all large and complex systems.

Agile Bi can solve many problems faced by enterprise bi. Agile Bi can sort, filter, and collect statistics for end users in a highly interactive manner without the guidance of Bi experts. With agile Bi technology, enterprises can enjoy the Value embodied in data more extensively.

Enterprise vs agile Big Data

Now is the time for big data technology to come out. EMC greenplum and teradata ASTER data are representatives of enterprises, while 1010 data and pervasive datarush are representatives of agility. Both vendors are aware of the gap between enterprise and agility in the Bi field and strive to solve this problem. Enterprise-level Big Data vendors know that they need agility, while agile Big Data vendors know that they need to provide high-quality enterprise-level solutions.

Enterprise-level big data providers have released some flexibility solutions. Greenplum launched a product named chorus to provide a collaborative environment that supports data analysis and search. Aster provides a series of SQL extensions that allow hadoop-type queries to use syntaxes similar to SQL. In terms of agility, the expansion of these products is greatly improved. However, people are more interested in the high price tags brought about by greenplum and Aster enterprise-level technologies: can we use agile big data technology to make up for the shortcomings of enterprise-level technologies and get more returns with less investment?

Agile big data solves the difficulties of enterprise-level platforms

Looking back at the three agile big data technologies we mentioned, the first question is: Why are they called agile?

The answer is actually very simple. They allow users to gain extraordinary data insights and cut prices. As follows:

• After some training, you can use splunk to query, filter, and display data.
• 1010data provides users with a big data processing interface in the form of workbooks
• Pervasive datarush processes data in parallel and efficiently on a GUI

Case study of agile big data processing in large batches

David INBAR is the Chief Executive Officer of pervasive's office of market development strategy. at a low price, he made an experiment demonstration to process a large enough data that can be called big data.

Datarush's toolkit eliminates the complexity of concurrent programming for programmers. Even tech geeks, parallel programming is nothing more than rocket science than writing a single-threaded program. Datarush provides a programming model for parallel programming to replace difficult multi-core function coding. For example, you can write a workflow application based on the component model. The toolkit automatically converts it to the parallel application allowed by the workflow.

Big Data Systems such as hadoop have horizontal scalability. Although datarush can assume such roles and can be installed on thousands of computers, the difference is that it is generally installed on one computer to explore the potential of multi-core systems.

The most distinctive feature of datarush is that it does not require you to know exactly how many cores a computer has. When writing a datarush application, it automatically senses the number of cores on the computer, and maximize utilization for parallel processing.

Integrate agile components to build an enterprise-level system

Agile technology has great advantages in creating intelligent business systems for big data, but there is still a long way to go. Agile Bi can reduce costs and prove its value in the face of Big Data Challenges.

From: http://www.searchsoa.com.cn/showcontent_53735.htm