DOWNLOAD MY Ph.D Thesis - UNAM

Chapter 6Optimisation of CRA using adaptive CSAs p6.2.2 Simple-CSA: Performance optimisation of the use of contentionslotsIn order to demonstrate the effects of the use of a simple contention slot allocator, anInternet traffic situation was analysed. The mean data rate per active station was set to64 kbps. The simulations were performed using both contention resolution algorithms.The Min. o. of CSs per signalling frame was ranged from 1 to 7 CSs for theexponential backoff algorithm. For the splitting tree algorithm this parameter wasranged from 0 to 6 CSs.6.2.2.1 Exponential backoff algorithm performance using a Simple-CSAWe have used Equation 4.43, (introduced in Section 4.4.3) to estimate the maximumupstream throughput for the exponential backoff algorithm.Smax_ exp_ backoffPksize=Pk + eCSsslotsThe term ‘eCSs’ in this equation, indicates that on average we need ‘e(=2.718)’contention slots to transmit a reservation request successfully, in order to get theoptimum system throughput, as suggested in [10]. Idles occur with a probability of 1/e ≈0.368, success occurs with a probability of 1/e and collisions occur with a probability of1-2/e ≈ 0.264. Thus in summary, 1 CSs for idle, 1 CSs for success and 0.718 CSs forcollisions are needed.On average the packet size for Internet traffic is 368.1 bytes [53], and the number ofdata slots requested per packet is 8.3. So using Equation 4.43 we have,Smax368.1== 52.2% ≈ 52%8.3*64 + 2.718*64Therefore, the maximum theoretical system throughput that can be achieved is ≈ 52%when large numbers of stations are transmitting.6-3