Enterprise network architecture and management have followed the Internet’s design principles despite different requirements and characteristics: enterprise hosts are administered by a single authority, which intrinsically assigns different values to traffic from different business applications. We advocate a new approach where hosts are no longer relegated to the network’s periphery, but actively participate in network-related decisions. To enable host participation, network information, such as dynamic network topology and per-link characteristics and costs, is exposed to the hosts, and network administrators specify conditions on the propagated network information that trigger actions to be performed while a condition holds. The combination of a condition and its actions embodies the concept of the network exception handler, defined analogous to a program exception handler. Conceptually, network exception handlers execute on hosts with actions parameterized by network and host state. Network exception handlers allow hosts to participate in network management, traffic engineering and other operational decisions by explicitly controlling host traffic under predefined conditions. This flexibility improves overall performance by allowing efficient use of network resources. We outline several sample network exception handlers, present an architecture to support them, and evaluate them using data collected from our own enterprise network.