The big data phenomenon brings new challenges to the world of data modeling. There are so many things in the NoSQL world that just don't fit into entity–relationship (E-R) or star-schema models. The first challenge is modeling to understand and describe existing data instead of to design databases, but that is the reality of big data—it exists without models and usually without sufficient metadata. We are also challenged by data constructs that don't fit neatly into E-R structures, including multivalued attributes, embedded arrays, nested documents, and associations implemented without foreign key relationships. In addition, key–value stores, document stores, and graph databases aren't easily described with E-R modeling. Clearly with big data, we need a new approach to data modeling.
E-R modeling is mature for SQL databases, but doesn't translate well to many of the NoSQL databases. This interactive workshop will introduce the Concept and Object Modeling Notation (COMN, pronounced "common"), which can represent the new structures NoSQL DBMSs support, while still supporting traditional SQL DBMSs. It also has more expressive power for modeling the real-world entities that databases are about. The workshop concludes with a realistic exercise in modeling a NoSQL operational database and a SQL data warehouse for a small enterprise.
You Will Learn
- New ways of thinking about the real world, data, and objects
- Modeling notation for containment and multivalued attributes
- How to model a variety of data, and NoSQL and SQL databases
- How to model things in a business-friendly way and how to map those things to data
- Data modelers, data architects, big data analysts, data scientists, and others looking for ways to visualize the structure of data in all its variety; people who need to design NoSQL and SQL databases; people looking for ways to graphically represent business requirements in preparation for modeling the data that represents them