Session initiation protocol (SIP) control traffic routing decisions, such as rapid failure detection and recovery mechanisms, are based on detection of events and/or conditions that may result in suboptimal performance. The disclosed approach ensures that the SIP traffic is routed or rerouted on an optimal basis. Sample inputs to the SIP routing decisions, include integration with BFD, system metrics and other criteria to determine network and systems conditions, for dynamic decisions on where to optimally route SIP traffic. Examples leverage the B2BUA model, which maintains call state and participates in call processing, however, the examples introduce an improved routing model that is able to not only reroute SIP traffic based on failures but also upon the detection of other suboptimal conditions, e.g. excessive cost. The exemplary techniques provide significant savings by ensuring SIP peering decisions are automatically performed considering costs of peering arrangements, congestion and/or other criteria.