Performance Analysis Of The Sleep Mode In IEEE802.16M With A Fixed Sleep Cycle

The Mobile Stations (MSs) of the broadband wireless access system use the batteryas an energy source. The energy consumption of MSs has been a key factor in restrictingthe application of the wireless broadband network. In order to reduce the energyconsumption and to extend the lifetime of the battery-powered MSs, WiMAX2provides anovel sleep mode scheme. Based on the mixed traffic of the real-time and the best-effort, asystem mathematical model is developed. With numerical results, the system performanceis investigated and the sleep mode in WiMAX2is optimized.Firstly, based on the working principle of the sleep mode in WiMAX2with themixed traffic, the time length of the sleep cycle is assumed to be fixed. Considering theextended listening window can end at any instant during a sleep cycle, the time length ofinitial sleep window can be regarded as one half of a sleep cycle. The subsequent sleepwindow is assumed to be a sleep cycle. A multiple heterogeneous vacation queueingmodel is therefore developed.Secondly, considering the correlationship of the data packets in the mixed traffic, thearrival of a data packet is assumed to be a Discrete Time Markov Arrival Process(D-MAP). Based on a limited capacity, a D-MAP/G/1queueing model with multipleheterogeneous vacation is built. By using the method of an embedded Markov chain, thequeueing model is analyzed. The expressions of the queue length and the busy cycle instead state are derived.Thirdly, considering the general random nature of data packets in the mixed traffic,the Phase-Type (PH) distribution is introduced to describe the input process of the systemmodel. A PH/Geom/1queueing model with multiple heterogeneous vacation is thenconstructed. Taking into account both the correlationship of data packets in the mixedtraffic and the general random nature of data packet transmission time, a D-MAPdistribution is introduced to capture the system input, and a PH distribution to representthe transmission process. A D-MAP/PH/1queueing model with multiple heterogeneousvacation is therefore constructed. By employing the method of matrix-geometric solution,the queueing model is analyzed, and the queue length in steady state is given. At last, the performance measures of the sleep mode, in terms of the average latencyof data packets and the energy saving ratio are derived. Numerical results are provided toinvestigate the influence of the system parameters on the system performance. Thetrade-off between different performance measures is revealed. A cost function is thendeveloped to determine the optimal length of the sleep cycle for the sleep mode inWiMAX2.