Exploring the path of wisdom development from data mart and integration

IBM's business intelligence began with data marts, helping IBM address reporting and http://www.aliyun.com/zixun/aggregation/11009.html "> Analysis issues for specific business areas such as finance and marketing." However, we also find that data marts create problems at the same time, such as the ability to achieve compliance, security, governance, and increasingly popular (increasingly scary) data redundancy! Many CFOs have been jailed for failing to disclose the truth, so we must be cautious about compliance and governance.

Evolution from data marts to enterprise data warehouses

To address the challenges of data marts and data sprawl, IBM and its rivals gradually began to focus on building enterprise data Warehouses (EDW). We built a consolidated, monolithic infrastructure to truly achieve a "single fact source." We integrate all the fields into the EDW, and the concept of the virtual field leads to new technologies that enable us to implement functions similar to the cube in the database engine.

We want to take control of the data, so we try to consolidate it, but there are other problems with keeping control. Maintaining high performance in these systems is particularly difficult, especially for some analysis workloads. Smaller vendors in the market have built a cheaper, software-like solution to handle analysis workloads, eliminating this functionality in EDW. These software devices are highly attractive to senior management in the pursuit of high-speed, economical solutions. Around 2003, such demand triggered a boom in such software devices in the marketplace.

But the data mart is drifting!

Ten years later, after all the lessons, we saw the trend of the data mart rising again. Have we found a way to solve the original problem of the data mart? Perhaps it should be said, in a way, but we must make sacrifices and compromises. One compromise is control and data, but we find that EDW fails to meet certain business requirements, such as ROI and rapid value-creating processes. I've seen up to 145% ROI data--not a joke. If you are the CFO, would you choose a dream ROI or a EDW?

Why does the CFO insist on choosing a data mart

Yes, I know the Data mart is in vogue for a while, but the reason why it is popular is that I cannot understand it. We built these excellent EDW structures, but why do we still fail in business? Why don't they focus on governance and compliance? If we do not manage the governance of these data marts, the dilemma of 20 years ago will reappear.

But it should not be forgotten that we can now say that the Data Warehouse has been successful because the pressure needed to better meet the business is more severe than ever before. This is really a riddle, isn't it? How do you manage the tradeoff between data governance and business demand responsiveness?

The development of wisdom: lessons learned from data marts and EDW

I think "putting all your eggs in one basket will achieve your goal" is by no means the right answer. Of course, it's simple and sounds reasonable-but it's only theoretical. Oracle is still implementing a one-size-Fits-all strategy, which has made me realize that Oracle's maturity level in data warehousing is far below IBM's. But this is not unacceptable. They will keep pace with IBM.

In fact, the monolithic architecture can solve some problems, but it will create some new problems. We have addressed some of the governance, metadata, security, and compliance issues. But our modeling and methodologies create some complexity and are therefore not recognized by business supporters.

In short, we have increased the pressure on the architects and the DBA team, resulting in a disconnect from the business. You must have a more realistic idea of how to implement a data warehouse.

To solve immediate problems, you must be able to take full advantage of consolidation while meeting the requirements of accelerating the process of creating value, which should be the case for analytical applications. You must align your business needs with the correct computing technology.

IBM's evolving EDW business strategy

In past articles, I've looked at IBM's new business strategy, which evolved from lessons learned from our data marts, EDW, and 20 business intelligence experience.