Your company needs to design a database that shows how changes traffic in one area of a network affect other components on the network. Which type of data store should you use?

Key/value stores are optimized for fast lookups by a unique key and have no native mechanism to model or traverse relationships between entities.

Graph databases store data as nodes (entities) and edges (relationships), enabling efficient traversal of how one network component affects others. This makes them the natural choice for modeling network topologies, dependency chains, and impact analysis - queries like 'which devices are affected if this router fails' map directly to graph traversal algorithms.

Document stores organize semi-structured data (e.g., JSON) and are suited for flexible schemas and hierarchical data, not for querying relationship propagation across a network.

Columnar stores are optimized for analytical aggregations over large datasets organized by column, not for traversing entity relationships or modeling network topology.