Archive for October, 2010

What Is a Data Cloud?

October 29, 2010

“Cloud” is the latest buzzword in our industry. We have heard of cloud computing, cloud services, etc. But I had never seen the phrase Data Cloud until recently. The concept is very simple. Find a way to link all the data islands under application silos. Therefore you need to have access to metadata and create graphs of relationships. There is a French company (headquarters in Novato, California) called Radiant Logic which offers a set of tools to pull data and virtualize. At a high level, it’s liberating the information and making it available for the business users.

The founders of Radiant Logic came from the world of directory and LDAP. The various layers would include ontology and glossary (data dictionary or repository) where common entities and attributes would be collected. Then you need systems of reference (source) via identified hubs or registry. Here the phrases CDI/MDM are loosely used. Then you need common federated query, also called virtualization. Some BI solutions address this in a limited sense mostly for retroactive analysis. Finally, you need search and browse capabilities using Xquery or an inference engine. This stack of various functions can yield what is called a true “data cloud”. Remember, this is not process integration. It is leaving data wherever it exists and creating some aggregation and consolidation for the user.

Then the question is where does such a “data cloud” fit into the lexicons of SaaS, PaaS, or IaaS? The vendors of data cloud claim it belongs to another form of SaaS delivery, where all the above functionality can be hosted in the cloud for users.

I will be very curious to see the actual deployment of such tools. This can be viewed as an advanced form of “data integration” we talked about years back.  So Radiant Logic provides context management and virtualization. Mark Logic provides XML and XQuery server. Thetus provides semantic knowledge modeling. A Burton Group analyst cited these three companies as those addressing the “data cloud” space.