It is now expected that many homes contain at least one Internet of Things (IoT) device. These devices have access to sensitive data, and information while operating within our homes. As IoT becomes more common place we must ensure that the users of these devices are protected. In general egress traffic from IoT devices is assumed to be intended by the manufacturer and is allowed by default. This paper explores the creation and enforcement of network policies that represent the expected behaviour of a given IoT device. We look at the potential of various technologies to allow low level observation, and filtering of egress traffic on the device itself. The information in the paper is largely theoretical, and focused on the security risks and benefits of specific methods. We propose a kernel level solution that could provide performance and security benefits in the future.