Graph Connectivity Explained: From Components to Traversals

What is graph connectivity?

Graph connectivity describes how vertices (nodes) in a graph are linked by edges. A graph is:

Connected (undirected): there’s a path between every pair of vertices.
Disconnected: it splits into two or more connected components — maximal subgraphs where every pair of vertices is mutually reachable.

For directed graphs:

Strongly connected: every vertex can reach every other via directed paths.
Weakly connected: replacing directed edges with undirected ones yields a connected graph.
Strongly connected components (SCCs): maximal sets of vertices with mutual reachability following edge directions.

Why it matters

Connectivity underpins many problems: network resilience, routing, clustering, influence spread, and verifying whether algorithms can traverse entire datasets.

Key concepts

Connected component: maximal set of nodes mutually reachable.
Bridge (cut-edge): an edge whose removal increases the number of components.
Articulation point (cut-vertex): a vertex whose removal increases the number of components.
Biconnected component: a maximal subgraph with no articulation points; every pair of vertices has two disjoint paths between them.
Reachability: can node A reach node B?
Giant component: in random graphs, a component containing a finite fraction of nodes above a connectivity threshold.

Basic algorithms

Depth-First Search (DFS) / Breadth-First Search (BFS): find connected components in O(V+E).
- Run DFS/BFS from an unvisited vertex, mark all reachable nodes as one component; repeat.
Kosaraju’s algorithm: find SCCs in directed graphs using two DFS passes (O(V+E)).
Tarjan’s algorithm: single-pass DFS to compute SCCs and articulation points/bridges (O(V+E)).
Union-Find (Disjoint Set Union): maintain and query components dynamically; useful for offline connectivity and Kruskal’s MST (amortized near-constant per operation).

Complexity notes

Most fundamental connectivity checks run in linear time O(V+E). Dynamic connectivity (online edge insertions/removals) requires more advanced data structures (link-cut trees, Euler-tour trees) with polylogarithmic update/query times.

Practical tips for implementation

Use adjacency lists for sparse graphs, adjacency matrices for dense graphs or fast connectivity checks with matrix operations.
Iterative BFS/DFS avoids recursion limits in large graphs.
For undirected graphs, mark edges carefully when detecting bridges/articulation points.
For large-scale or streaming graphs, prefer Union-Find or dynamic-tree structures.

Example use cases

Network failure analysis: identify bridges and articulation points.
Social network analysis: detect communities via components and SCCs.
Compilers and program analysis: SCCs for detecting cyclic dependencies.
Geographic routing and map connectivity: verify reachable regions.

Graph Connectivity Explained: From Components to Traversals