Introduction to Cassandra

Overview

• Cassandra: A distributed NoSQL database known for high availability and scalability.
• Key Topics Covered: Features of Cassandra, Read/Write Path, Data Modeling, Use Cases.

Features of Cassandra

• Distributed Database: Data is replicated across multiple machines.
• Always Available: Works even if one or more machines fail.
• Eventual Consistency: Tunable to strong consistency if needed.
• Fast Writes: Suitable for write-intensive applications.
• Leaderless Architecture: No single point of failure; all nodes can handle writes.
• Peer-to-Peer: More resilient than master-slave setups in relational databases.

Distributed Nature and Availability

• Data Replication: Same data resides on multiple nodes for redundancy.
• High Availability: As long as one replica node is available, queries can be fulfilled.
• Resilience: Any node can take over the role of a failed node.

Storage and Data Structure

• Partitions: The most atomic unit in Cassandra; akin to files within a system.
• Designing Partitions:
  • Related data should be in the same partition.
  • Smaller partitions lead to better query performance.
• Schema Design:
  • Needs to be carefully planned with query patterns in mind to maintain performance.
• Example Schema:
  • Single Field Primary Key: Defines one field as the partition key.
  • Composite Partition Key: Includes multiple fields.

Write Path Efficiency

• Memory-Based Writes: Data first goes to a commit log and memtable (both in memory).
  • Commit Log: A pen-only log.
  • Memtable: A data structure in memory.
• Speed: Writes are acknowledged once data is in memory, not waiting for disk storage.
• Disk Storage: Data is later flushed from memory to disk (ssTables) periodically.

Use Cases

• High Write Throughput: Ideal for applications requiring frequent and fast writes.
• Internet of Things (IoT): Suitable for high-frequency sensor data.
• Web Activity Tracking: Efficient for logging user interactions on busy websites.
• Predictable Read Patterns: Highly efficient if queries are defined ahead of time.
• Health Data Applications: Fast write performance supports constant health monitoring data.

Conclusion

• Cassandra excels in resilience, availability, and write performance.
• Ideal for applications with high write requirements and predictable read patterns.