Packet-level QoS over RPL
Peter Thai, Jaudelice Cavalcante de Oliveira · 2011
Low-power and Lossy Networks (LLN) are composed of embedded devices constrained by computation capability, memory capacity, communication bandwidth, energy supply, and also node and link reliability. The IETF ROLL (Routing Over Low-Power and Lossy networks) Working Group has specified RPL (Routing Protocol for Low-Power and Lossy networks) to address these constraints, after evaluating other protocols proposed in the literature in search of an IP based solution for routing over LLNs. The positive preliminary results of RPL for providing a reliable LLN for a single application naturally leads to the discussion of multiple applications sharing network resources. An immediate issue is the fair and efficient allocation of the shared wireless channel bandwidth. These applications may require specific network guarantees for system critical warnings, time-sensitive data, and general data of less urgency. Packet-level Quality of Service (QoS) has been a very popular paradigm for ensuring bounded delay and maximum network throughput for well-behaved flows in wired and ad-hoc networks, and to the best of our knowledge, no comprehensive efforts to document traditional QoS performance over LLNs exists in the literature. In this thesis, we summarized the motivations for the development of RPL, briefly describe it and then discuss traffic classification in LLNs. We present the simulator that was developed to evaluate RPL and the extensions needed to support traffic classes and scheduling mechanisms. Finally, we evaluate a very simple QoS scheduling discipline using data collected from a real smart metering network deployment, indicating the feasibility of supporting QoS for such networks.