- Relational databases
- NoSQL databases
- NewSQL databases
- Time-series
-
Key points
- Relational databases are based on the relational model, which organizes data into tables with rows and columns.
- Relational databases is suitable for storing structured data.
-
Pros
- Structured data
- Data in relational databases is stored in tables with a predefined schema, enforcing a consistent structure throughout the database.
- Relationships and referential integrity
- The relationships between tables in a relational database are defined by primary and foreign keys, ensuring referential integrity.
- SQL support
- Relational databases use Structured Query Language (SQL) for querying, manipulating, and managing data. SQL is a powerful and widely adopted language that enables developers to perform complex queries and data manipulations.
- Transactions and ACID properties
- Relational databases support transactions, which are sets of related operations that either succeed or fail as a whole. This feature ensures the ACID properties are maintained, guaranteeing data consistency and integrity.
- Indexing and optimization
- Relational databases offer various indexing techniques and query optimization strategies, which help improve query performance and reduce resource consumption.
- Structured data
-
Cons
- Limited scalability
- Scaling relational databases horizontally (adding more nodes) can be challenging.
- Low flexibility
- The predefined schema in relational databases can make it difficult to adapt to changing requirements, as altering the schema may require significant modifications to existing data and applications.
- Performance issues
- As the volume of data grows, relational databases may experience performance issues, particularly when dealing with complex queries and large-scale data manipulations.
- Inefficient for unstructured or semi-structured data
- Relational databases may not be suitable for managing unstructured or semi-structured data, such as social media data or sensor data.
- Limited scalability
-
Common products
- MySQL
- PostgreSQL
- Oracle Database
- Microsoft SQL Server
- Amazon RDS
- SQLite
- IBM Bb2
- MariaDB
-
Key points
- NoSQL databases can store data in various formats, which makes them suitable for a diverse range of use cases.
- NoSQL databases is suitable for storing semi-structured data.
-
Sub-types
- Document-oriented
- Concept
- Documents encapsulate and encode data in some standard formats or encodings (JSON,XML,YAML,BSON).
- Pros
- Flexibility (Schemaless)
- Use cases
- Need to store hierarchical or nested data (JSON, XML).
- Products
- MongoDB
- CouchDB
- Terrastore
- OrientDB
- RavenDB
- Concept
- Column-oriented (Wide-column)
- Concept
- Organize data by columns rather than rows.
- Pros
- Provides improved compression and better read performance.
- Use cases
- Need to store and query large amounts of data across many nodes (Popular choice for big data and analytics).
- Need to handle high write and read workloads.
- Products
- Cassandra
- HBase
- Hypertable
- Amazon SimpleDB
- Concept
- Key–value
- Concepts
- Store data as key-value pairs.
- Pros
- High read and write performance.
- Horizontal scalability.
- Use cases
- Needs to handle high-speed reads and writes
- Caching layers
- Session stores
- Configuration storage
- Needs to have low-latency access to data
- Gaming platforms
- Real-time analytics systems
- Recommendation engines
- Needs to handle high-speed reads and writes
- Products
- Redis
- Memcache
- Amazon DynamoDB
- Cassandra
- Couchbase
- Concepts
- Graph
- Concepts
- Store data as nodes and edges in a graph
- Pros
- Efficient processing of complex relationships, traversals, and graph-based algorithms.
- Provide powerful querying capabilities for traversing and analyzing interconnected data.
- Use cases
- Need to involve intricate relationships between entities.
- Social networks
- Fraud detection systems
- Recommendation engines
- Need to involve intricate relationships between entities.
- Products
- Neo4J
- Amazon Neptune
- Infinitegraph
- OrientDB
- FlockDB
- Concepts
- Document-oriented
-
Cons
- Lack of standardization
- NoSQL databases often use their own query languages or APIs. This can lead to increased learning curves and difficulties.
- Weaker consistency
- Many NoSQL databases employ eventual consistency models to achieve higher performance and availability.
- Limited support for complex queries and transactions
- Some NoSQL databases are not designed for complex queries or multi-record transactions.
- Lack of standardization
-
Key points
- Modern approach to combining the strengths of both relational and NoSQL databases.
- They maintain the relational model, ACID properties, and SQL support.
- They offer improved scalability, distributed architecture, and performance enhancements.
- Modern approach to combining the strengths of both relational and NoSQL databases.
-
Pros
- Distributed architecture
- NewSQL databases are distributed. They leverage data partitioning and replication across multiple nodes or even data centers. This architecture allows for better fault tolerance, high availability, and global scale.
- Horizontal scalability
- NewSQL databases can scale horizontally (Adding more nodes to the system).
- Concurrency control
- NewSQL databases use advanced concurrency control mechanisms (e.g. multi-version concurrency control (MVCC) or optimistic concurrency control). These mechanisms allow efficient handling of a large number of simultaneous transactions.
- SQL support
- NewSQL databases support SQL for querying and manipulating data.
- Simplify the learning curve for developers.
- Provide compatibility with existing relational databases and tools.
- simplify the migration process.
- NewSQL databases support SQL for querying and manipulating data.
- Distributed architecture
-
Cons
- Complexity
- NewSQL databases can introduce additional complexity in terms of configuration, maintenance, and troubleshooting.
- Vendor lock-in
- Some NewSQL databases are offered as managed services by specific vendors, which may lead to vendor lock-in and limit the flexibility to switch providers.
- Lack of maturity
- NewSQL databases may lack the maturity and extensive ecosystem.
- Complexity
-
Products
- Google Cloud Spanner
- CockroachDB
- NuoDB
- VoltDB
- TiDB
-
Key points
- Specialize in handling time-stamped (time-series) data.
-
Pros
- High write and query performance
- Time-series databases are optimized for handling high-velocity data streams, which require efficient write performance.
- Time-series databases provide fast query performance, allowing for real-time or near-real-time analysis of time-series data.
- Data compression
- Due to the large volume of data generated by time-series workloads, time-series databases use various data compression techniques to reduce storage requirements.
- Data retention policies
- Time-series databases enable easy management of data retention policies based on time. This feature allows for automatic data aging, which helps to maintain storage efficiency and ensure data relevance.
- Built-in time-series functions
- Time-series databases typically include built-in functions and tools to facilitate time-series data analysis, such as aggregation, downsampling, and forecasting.
- Horizontal scalability
- Time-series databases are designed to scale horizontally, allowing them to handle large volumes of data and high ingestion rates.
- High write and query performance
-
Products
- InfluxDB
- TimescaleDB
- Prometheus
- Graphite
- Relational and NoSQL database
| Relational database | NoSQL database | |
|---|---|---|
| Schema | Fixed | Dynamic |
| Query Language | SQL | UnQL (Unstructured Query Language) |
| Scalability | Vertically scalable | Horizontally scalable |
| Transaction Guarantee | ACID | BASE (for performance and scalability) |
| Sub-types | ||
| When to use |