RSS
热门关键字:  数据挖掘  人工智能  数据仓库  搜索引擎  数据挖掘导论

Metadata Management

来源: 作者:unkonwn 时间:2004-12-03 点击:

Whenever you ask an IT person to list the ten major issues in their organization, metadata is always on the list and always about number nine or ten. Every year, some new issue rises to number one, and metadata perpetually stays at nine or ten. It is the eternal “wait until next year” issue. As a consequence, metadata problems often seem to be ignored. 数据挖掘研究院

There are multiple reasons for the failure to address metadata issues. Many of the products for metadata management were developed during the early decades of the industry. Consequently, some of these products are solving yesterday’s problems, not today’s problems. Additionally, the track record for metadata has not been sterling. Organizations remember that they spent a lot of money and energy on data dictionaries and centralized repositories, only to have those projects fail. Then there is the issue of a metadata budget. The budget is usually allocated to hot projects, not metadata.  数据挖掘研究院

Poorly managed metadata is a symptom of a larger disease. When there are legacy systems that reflect yesterday’s requirements, getting the metadata in order is somewhat like arranging the deck chairs of the Titanic. Compounding the problem is the fact that architecture has changed. In the early days, there was a centralized mainframe architecture. Now there are distributed systems – data warehouses, operational data stores, data marts, virtual operational data stores, and so forth. The architecture of the past has changed dramatically while the concepts and technologies surrounding metadata have not. In order to be effective, it is necessary to have proactive metadata management – not passive, optional metadata management. 

数据挖掘研究院

There is a crisis point that is being reached, and the world desperately needs a good metadata solution. The forces driving the need for a viable metadata solution include:  数据挖掘研究院

  • The volume of data continues to grow at an exponential rate.

  • As data ages, there is a greater need for better metadata management.

  • There are two different kinds of metadata that need to be managed: technical and business.

  • The structured environment must now address different kinds of data such as e-mail, documents, letters and even telephone conversations.

  • Architectures are becoming even more distributed. 

The first thing that a good metadata solution will exhibit is an acute awareness of the adherence to the value proposition. No longer can metadata be sold as a technology. Metadata must be presented as a vital, business-aware, essential component of the corporate environment. If a metadata initiative is to succeed, the importance of metadata must be sold to the business person, not the technician.

数据挖掘研究院

Perhaps metadata needs to be sold as something other than metadata. Perhaps the problem is that the term metadata has too much negative baggage associated with it. However we do it and whatever we call it, it is clear that the time for a savvy metadata solution is now. 数据挖掘研究院

最新评论共有 0 位网友发表了评论
发表评论
评论内容:不能超过250字,需审核,请自觉遵守互联网相关政策法规。
匿名?