Preferential attachment → P(k) ~ k^{-3} power-law degree distribution
Barabási-Albert (1999): new node connects to m existing nodes with prob P_i ∝ k_i (preferential attachment).
Produces scale-free network with P(k) ~ k^{−3} for large k, independent of m.
Hubs = highly connected early nodes. Found in WWW, protein networks, citation graphs.