Real-Time Diameter Monitoring For Time-Evolving Graphs

The goal of this work is to identify the diameter, the maximum distance between any two nodes, of graphs that evolve over time. This problem is useful for many applications such as improving the quality of P2P networks. Our solution, C-Scale, can track the diameter of time-evolving graphs in the most efficient and correct manner. C-Scale is based on two ideas: (1) It estimates the maximal distances at any time to filter unlikely nodes that cannot be associated with the diameter, and (2) It maintains answer node pairs by exploiting the distances from a newly added node to other nodes. Our theoretical analyses show that G-Scale guarantees exactness in identifying the diameter. We perform several experiments on real and large datasets. The results show that G-Scale can detect the diameter significantly faster than existing approaches.