Theoretically data architecture can be defined as the compilation of concepts with regard to data, such as data model, access methods of data, strategies of data access, sustaining mechanism, data recovery and backup process. In order to put together an appropriately defined methodology, any organization should have apparent understanding of its data architecture system. 

With my experience all the way through my career, I had the opportunity to work under a number of data architectures. As an example, when I was holding the responsibility for developing “risk management system”, I had to realize that the data model the “clearing and settlement system” which is absolutely put together by diverse groups for   specific purposes. Therefore,  I had to incorporate “ risk management system” together with the above system.  Data model of the two systems were dissimilar, although in fact, we had to build interface to be compatible with each other in order to provide inter connected data transfers between the two systems. As an example  a “clearing firm” can be defined in the both systems by sharing  set of common  characteristics  and during the  operational period  each system should have a common  perception of the “clearing firm“  while  maintaining their data in two  different  settings  and endow with different functionalities in  each system. 

The “clearing system” was subject to absolutely diverse set of strategies, access mechanism, backup and maintains mechanisms and risk management system.  On the other hand, the “risk management system” depends on data architecture of the “clearing” system while it has its unique data architecture. 

Data modeling

Data modeling can be defined as putting together a model in order to represent the structure of data and relations within the scope of a specific project.  Data modeling is a critical part of any information system, since it offers a conceptual form which is not complicated to realize and easy to communicate with other parties. 

Most of the time, at  the beginning of the project,  we endeavor to put together a data model for the organization by using a variety of modeling concepts such as  entity relational model or object model. There are various perspectives of the model depending on the circumstances. High level representation of significant entities is identified as conceptual data model.  Foremost objective  of this is to build overall picture of the data model rather than building detailed model. More detailed representation of data model with complete attributes, inter relations between entities can be identified as logical data model. The physical data model is then defined to represent the actual physical form of the system, like database instances, table space, and CPU, memory and storage necessities.  

Entity Relation diagram

ER diagrams are the most extensively used modeling concept of database designing and it has rich set of features which make available easy to realize design, even for comparatively non expert users, since of its diagrammatic approach. ER diagrams primarily consist of entities, characteristics, and relationships between entities. Entity may represent a thing in authentic world or can be a concept.  As an example “trading firm”, “user’, “trading session” can be represented as entities. Entity can be defined with set of characteristics, for above example “user’ entity may have attributes such as “name”, “address”, “age” etc.  A relationship is created when several entities are  put together in an association with each other. As an  example  the statement a “user’ is registered with a “trading firm” represents a relationship between “user” entity and “trading firm” entity.