Data Architecture

One must first understand the existing Enterprise Data Architecture Landscape as every new project’s data architecture must follow existing Landscape rather introducing new components with every new project. I have personally seen organizations creating Data Components/ Tools Swamp, as in current era of Big Data there are hundreds of tools offering the same functionalities so one must know whether a tool is already there in existing Landscape or not for a particular requirement. I would consider this as, one the biggest Risk in current era.

Data Architecture is the most important aspect of any data related solution or project. As mentioned in the previous chapter, there are four types of architecture i.e., 1) Business Architecture 2) Data Architecture 3) Application Architecture and 4) Technical Architecture. Most of the organizations confuse or mix-up the last three, which is again a recipe for disaster. Data Architecture should remain standalone from Application or Technology Architectures.

Data Architecture is driven by Data Architects and work towards achieving business value, compliance and implementation process aligned with Enterprise Data Architecture.

Enterprise Data Architecture consists of many components but two of these are very critical i.e., 1) Enterprise Data Model and 2) Data Flows.

1. Enterprise Data Model includes conceptual, logical, and physical data models with all the possible data-points across the organization including entities, attributes, relationships, and rules. It must be agreed by all the stakeholders, before all upcoming projects.
2. Data Flows represent the data movement across the organizations either via databases, applications, networks etc. Data Flows define the relationship between entities and business processes e.g., a Product is created by Product Development Process, produced by Manufacturing Process, sold by Order Management Process etc.

Steps for PROJECT/ PRODUCT Data Architecture

Below are steps to start working on Data Architecture with respect to a particular project or solution or product. Please note, the data architecture of every project must align with Enterprise Data Architecture.

• Outline Architectural Scope
• Business Requirement Scope
• Design and Plan
• Implement

Implementation Methodologies

We all know there are two famous implementation methodologies. In the implementation step, one must decide which methodology project is going to follow.

• Waterfall follows sequential implementation approach where each phase or module is dependent on previous one.
• Agile follows an object-oriented approach where multiple activities can be started in parallel using Sprints. DevOps, DataOps, MLOps all are inherited from Agile methodology.