IT Architecture: Data Layer

Without data an application will not work. Data is provided as input or transferred from other sources. 

 

There are 3 main types of data. 

1. Master data is information that does not change with every transaction. For instance, a client’s name or address, a product price is typically static without frequent changes. These entities are the main data objects in an application and are required to do transactions.
2. Transactional data records the purchase, sale, calculation and involves master data entities. For instance a ticket sale involves a client entity (who buys the ticket), his/her credit card number, the specific origin and destination route. The transaction is the specific sale of the ticket with date, hour and final price.
3. There are also reference data objects, those are data objects that can be used to describe in master or transactional data, and are typically pre-determined values. Common reference data sets are countries, currencies, car model, type of identification. 
4. Unstructured data: these data types above are all structured data, that is information that can be typed in, i.e. one can write the information with a keyboard. But there is also unstructured data which is information that cannot be typed in and mostly refers to images, videos other types of multimedia.

Data is typically stored in tables with rows and columns. If these tables can be related to each other, they are called relational data. This kind of data is usually stored in classic database technologies such as SQL, Oracle or other products.

Non-relational data is where there is no direct relationship or dependency between the data sets or data objects. In this case, to store one needs to use non-relational databases that use a storage model that is optimized for specific requirements of the type of data being stored. 

What is key of data design is that the data structure is well-designed in a way that it is efficient from an accessibility point of view as well as performance and scalability.