Relational databases can store large quantities of data, but they become less suitable when data has extremely high volume, velocity, or variety, or lacks a consistent structure. Their main limitations are rigid schemas, difficult distributed scaling, and reduced performance when complex joins must process enormous datasets.
A relational database organizes data into tables containing rows and columns. A predefined schema specifies each table’s fields, data types, keys, and relationships, supporting validation, consistency, and efficient structured queries using SQL.
However, unstructured data such as videos, images, audio, emails, and free-form documents does not naturally fit fixed rows and columns. The database may store such files as binary objects or store links to them, but it cannot easily query their internal meaning without additional processing.
| Limitation | Why it matters |
|---|---|
| Fixed schema | Changing fields or data types may require schema migration. This is inefficient when data structures change frequently. |
| Horizontal scaling | Distributing related tables across multiple servers is complex because queries, transactions, and relationships must remain coordinated. |
| Expensive joins | Combining huge tables can require substantial processing time, memory, and network communication. |
| Unstructured-data mismatch | Media and free-form text cannot be represented or searched as naturally as structured records. |
| Transaction overhead | Maintaining ACID properties improves reliability but can reduce speed and availability in very large distributed systems. |
This does not mean relational databases cannot handle big data. Powerful hardware, indexing, partitioning, replication, and distributed SQL systems can extend their capacity. However, NoSQL databases may be more appropriate when flexible schemas, rapid horizontal scaling, or specialized storage for documents, key-value pairs, or graphs is required.
For an IB Computer Science A3.1 response, identify a limitation and explain its consequence. Avoid the misconception that relational databases simply “cannot store big data.” A stronger answer states that they may become less efficient or harder to scale than alternative database models, depending on the data and application requirements.
