News Stay informed about the latest enterprise technology news and product updates.

Real-time data warehousing

The relationship between the local and global ODS.

This article originally appeared on the BeyeNETWORK.

For years in the corporate information factory there has been the operational data store (ODS). The ODS is the place where real-time data warehousing takes place. It is the place where sub-second transaction processing takes place. It is where direct updates can occur if needed. It is the place that satisfies the need for operational processing in the data warehouse.

In many ways the ODS is a hybrid structure. The ODS serves both the tactical need for OLTP processing and the managerial need for looking at a sea of integrated data in a high-performance mode.

There are two types of ODSs—the “global” ODS and the “local” ODS. The global ODS is the one that is found directly attached to the data warehouse. The local ODS is one that serves the immediate needs of the web-based business environment.

While there is almost always just one global ODS, there may be many local ODSs.

In order to understand the differences between a global ODS and a local ODS, let’s use the information needs surrounding a customer known as “Mike Laper,” as an example. In the global ODS there is a great deal of information about Mike that the corporation may need to know, such as:

  • Name;
  • Physical address;
  • When first purchase was made;
  • How much total sales amounted to;
  • Employer;
  • Occupation;
  • Social security number;
  • Bank (checking account); and
  • Credit worthiness, etc.

The information found in the global data warehouse is useful to many parts of the organization.

Suppose the organization has a hobbies’ website. On the hobbies’ website there are many items for hobbies that are for sale. The website has coin collecting, stamp collecting, dolls, trains, darts, golf, hunting and a wide variety of equipment and accessories to suit the hobbyist.

The website also has an ODS and in that ODS there is a record for Mike Laper which might contain:

  • Mike’s html cookie, so that Mike’s visits can be recognized;
  • The last time that Mike visited the site;
  • The last purchase that Mike made;
  • The categories of hobbies for which Mike has shown interest;
  • The date Mike first came to the website; and
  • The last promotion sent to Mike, etc.

The information found on the local website is that information that is specific to and of interest to Mike as a hobbyist.

Of course, the local and the global websites can be connected. This is usually done by asking Mike to give some registration information that can be used to link him with his global ODS record.

Also note, that there may be other local ODSs that include Mike. Suppose the corporation has a music lovers’ website, in addition to the hobbyist website. On the music lovers website there may be another local ODS record for Mike. This local ODS might have such information as:

  • Mike’s html cookie;
  • Mike’s last CD purchased;
  • Category of music preferred by Mike—country and western, grunge rock, gospel, etc.;
  • Last music promotion sent to Mike; and
  • Last time Mike visited the music website.

If there is more than one local ODS, the data on the local ODS can be linked by matching cookies.

Also note that data can be exchanged between a local ODS and a global ODS bi-directionally. If it makes sense to exchange data, then there is no reason why such a transfer cannot be made.

The value of creating a global ODS is its usefulness in tracking a customer through various channels within the enterprise.


Bill Inmon is universally recognized as the father of the data warehouse. He has more than 36 years of database technology management experience and data warehouse design expertise. He has published more than 40 books and 1,000 articles on data warehousing and data management, and his books have been translated into nine languages. He is known globally for his data warehouse development seminars and has been a keynote speaker for many major computing associations.

Dig Deeper on Enterprise data architecture best practices

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.