Abstract
The unprecedented increase in the number of Internet users, routers, and service providers has introduced significant challenges to the design of scalable network architectures and end-to-end protocols. Web driven demand and traffic can exhibit extreme variability; providing predictable quality of service (QoS) without resorting to major over-provisioning is a difficult problem; facilitating dynamic group communication and multicast has spurred a multitude of proposals, each with its own idiosyncrasies and trade-offs; QoS routing faces the computational complexity barrier; congestion control is asked to be fair, efficient, and stable in a complex environment; mobility and wireless channels impose new control dimensions and constraints; and faults in software and hardware introduce disruptions that may persist in time and spread in space. A common denominator to many of these examples is scalability, which, to varying degrees, plays an important role when designing and evaluating feasible solutions.