The Network Oracle.

This paper sets out a high-level research agenda aimed at building a collaborative, global end-system monitoring and information infrastructure for the Internet’s core state. We argue that such a system is beneficial, feasible and timely, representing an important area for engagement across database and networking technologies. We start by hypothesizing the benefits to the Internet of a “Network Oracle” that could answer real-time questions about the global state of the network. We then argue that database and networking technology should make it possible to provide a useful approximation of such an oracle today, gathering information from a large number of end hosts and delivering useful views to each end system. We further argue that this can and should be done in a decentralized fashion. We provide an outline of such a system: it employs sensing agents, along with a distributed query/trigger/dissemination engine, and possible attractive end-user applications. A key point of our discussion is the timeliness and importance of a grassroots agenda for Internet monitoring and state-sharing. While significant social and economic barriers to deploying a centralized, public Internet monitoring infrastructure exist, there are corresponding social and economic incentives for a collaborative approach.