Large data provides predictive and analytical services for intelligent business

Chambers is vice president of IBM's Analytics Solutions division. She says many customers have enough to buy large data and predictive analytics services, but want them to implement the old business intelligence tools and database tools they have adapted to.

"Often customers do things that they already know. Chambers said at the 2012 Hadoop summit. "They want to take advantage of their existing infrastructure, use existing data and tools." They don't want any difference or change. So I told my client that if you don't do different things, you won't get any different results. ”

New methodologies, technologies and tooling requirements

Chambers is perfectly correct, as follows from three aspects of the analysis of the reasons:

1. Infrastructure. New approaches to processing and storing large, multiple-structured datasets are emerging because traditional relational technologies cannot be completed or cost effective in a unit of time. For example, Hadoop allows you to run open source software within a reasonable amount of time to store and process large data at very low prices. Now try to take advantage of Oracle's services. Can save 3 million of dollars in costs, as well as 6 months of time.

2. Data. Large data is rich in existing internal transaction data and other different sources of data, these sources are from outside your enterprise. This could mean that the data came from Twitter or http://www.aliyun.com/zixun/aggregation/1560.html ">facebook social media, or from the National Weather Service, Data from the public sector in the education sector, from Bloomberg and Dow Jones. If you don't have mashup data, you may not need to do large data analysis.

3. Tools. Most traditional business intelligence tools will not be cut, since they must be on a parallel computing infrastructure of new, larger, and more diverse volumes of data. What you need is a modern data visualization and analysis platform that allows users to easily handle large data visualization. To be fair, a handful of existing business intelligence providers, such as Tableau and microstrategy, are trying to better integrate their products into big data. In general, however, the old reporting tools that you have used in the past 10 years or so are unable to provide enough operational insights into the current large data.

The game of risk

But as far as I know, this change is difficult, so sometimes it is to avoid risk. But we are at a crossroads. Big data is by no means a flash in the pan or a slightly better way of business intelligence. This is a new paradigm that requires a major shift in thinking. In other words, "you're already going through some extra risk." "To achieve the success of big data, as Chambers say.

She says that means "if you want to have more insights, you have to inject new information into your application to your data network." "This means that you have to invest in new infrastructure technologies, such as Hadoop and other platforms, to form the basis for a new large data analysis." You need to use a new end-user tool to translate all the big data into Easy-to-understand insights.

The good news is that you don't have to undo your entire existing infrastructure and toolset. In fact, I strongly disagree with that. The business intelligence and data Warehouse you are using now may be a reason, because they are already providing you with the appropriate business value. In fact, many large data technologies can actually help you get more value from existing databases and tools.

When it comes to big data, start with small things. Identify a specific business problem that needs to be addressed, and a fixed business can bring tangible benefits. Exchange learning with peers in the big data industry.

(Responsible editor: The good of the Legacy)