On Wednesday I wrote about semantic technology's implication for the enterprise. In that post, I shamelessly hinted that it'd be great if the Cutter Consortium would offer non-members a look at the Cutter piece "The Rise of the Semantic Enterprise."
Guess what? Cutter did. And it turns out it's not one article, but a 40-page issue of the Cutter IT Journal, all devoted to semantic technology and its implications for the enterprise.
I'm still working my way through it, but I wanted to share the link with IT Business Edge readers, just in case it's a limited-time offer. If you have any interest in semantic technology, you'll want to read this journal.
Even if you don't give two figs about semantic tech, if you're investing in master data management or data integration, you definitely need to read at least one article, "Leveraging the Semantic Web for Data Integration." And here's why: This article explains how semantic technology can be used for integration and how the semantic queries could eliminate the need for major integration projects and master data management hubs.
To be fair, I should temper that statement by adding that I'm not 100 percent sure to what extent-entirely or just partially - the semantic queries would replace MDM and data integration. But the example given in the article is pretty darn impressive.
First, you have to understand a bit about SPARQL, a SQL-like query language recommended by the W3C and used for semantic queries. The writer, Shamod Lacoul, explains SPARQL can gather data from different URL locations:
Imagine a platform where the existing data, be it internal or external, is mapped to RDF (Resource Description Framework) and exposed via URIs just as you would expose Web services. Now, instead of calling the API, you can write queries against your federated data explicitly and merge the output using the same query. This provides a nimble way to embed business logic in the query by which you can fetch data from distributed data models that are rich, dynamic, and evolving.
What does this mean for the average company? This is where his example makes it clear that it could eliminate some complicated, time-consuming integration work and maybe even some tools:
For example, say your company has its customer data spread out among an SAP ERP server, an Oracle database, and SalesForce CRM. Previously, you would either have integrated all three systems to synchronize customer data among applications or used some type of master data management (MDM) hub to normalize data to a single version of the truth. Now, given that all customer data from the three systems can be exposed as RDF for access, you can simply SPARQL these disparate customer data sets and retrieve whatever customer information you need without much hassle.
The article goes into great detail about how the semantic data model, which is based on graphs, differs from the table-based format of relational database management systems and XML's tree-schema. Understanding that helps you to understand how semantic integration works.
The article also explains how semantic integration can be applied to internal integration, B2B integration and even the cloud. The discussion of Linked Open Data Cloud is particularly exciting, since that effort now includes The New York Times thesaurus data and GoodRelations, for defining company and products, as well as ontologies being created by the U.S. National Center for Biomedical Ontology and e-Government.
Much of the piece is very accessible, particularly in the early half of the article, but beware: There are some tech-laden sections that speak primarily to technologists and are a bit dense for those without an IT background.
Thank you to the Cutter Consortium, particularly Group Publisher Kim Leonard, for sharing this great resource.