Big data is everywhere. One increasingly important type of big data is in the form of networks: gene regulatory networks, disease networks, and social networks. Most of these networks are not static: the data constantly evolves according to some unknown but measurable dynamics. As the network behavior evolves over time, the resulting data becomes too large and comes in too fast to even store in the computer’s memory, requiring (sketching and sampling-based) algorithms (1) to manipulate the data immediately as it streams by using relatively little memory, or (2) to parallelize the data on a big number of machines having moderate memory by using few communication rounds.