Abstract
Much contemporary usage of the Internet is content-driven. The traffic originates from highly distributed services where users exclusively care about the data they are receiving, not where it comes from. However the longstanding architecture of the Internet and its management are largely host-oriented. Namebased networks address this dichotomy by proposing a content- and data-centric approach. In name-based networks, every interaction is driven by the name of the data that is being requested. Although this aligns with the large proportion of the way data is retrieved over the Internet, it remains an open question as to how this mechanism could be used to manage the forwarders in name-based networks and support autonomous operation. Specifically, in the context of a name-prefix based QoS mechanism, policies need to be present on every forwarder on the network, requiring policy distribution and continuous updates. Currently there is no standard solution to distribute and manage QoS policies on name-based networks. This paper explores how name-based networks could be managed autonomously, identifying mechanisms to control QoS policies across a network by examining an example scenario. We present a novel design for a QoS distribution mechanism that is conceptually coherent with existing name-based protocols; we subsequently illustrate this proposal by sketching out a concrete use case scenario.