Efficient measurements of IP level performance to drive interface selection in heterogeneous wireless networks

Several solutions have been proposed for mobility management in IP based heterogeneous networks, working at different protocol levels, from layer 2 up to application level. In order to take the handover decision, many solutions require to monitor the performance of the heterogeneous networks to which the mobile device is connected. Measuring the physical or link level performance on a given wireless access networks does not provide a reliable indication of the perceived level of service when the application flows are handed over that wireless access network. It is therefore needed to take measurements at IP level, on the (bidirectional) path from the Mobile Host to an intermediate node handling the mobility or even to the remote Correspondent Host. Gathering such measurements in a timely, effective and efficient way is not an easy task. In this paper we show that a naïve approach using application level active measurements is highly CPU intensive. This would severely impact battery usage in Mobile Hosts and does not scale if intermediate mobility management nodes are involved. On the contrary we show that an implementation of active measurements in the Linux kernel has a very low CPU usage. In this approach an efficient use of batteries in Mobile Host can be achieved and intermediate mobility management nodes can scale up to monitoring thousands of flows towards Mobile Hosts. Finally we discuss how combining passive and active measurements could further improve the solution.