Generating traffic Sample Clauses

Generating traffic. ‌ The Utilisation parameter for a test, specifies the target value for utilisation, i.e. ag- gregated throughput, on the backbone link. This value is an average for the generated backbone utilisation, taken over all runs (realisations) in a test. The traffic for each of the three service types is generated according to the traffic models, described in section 2.3.2 and appendix C. Based on the average throughput of each traffic type, which is shown in table 6.2 this results in an expected value h(s) for the throughput of a flow, for each of the service types s. That is h(s1)  7  h(s3) h(s) = h(s2) = 0.3 [Mb/s] (6.1) for s1 = video, s2 = voice, s3 = besteffort. For each run, i in a test, a random number of flows xi(s) are generated for each of the service types s, i.e. xi(s) = xi(s2) (6.2) xi(s1) xi(s3) for s1 = video, s2 = voice, s3 = besteffort. A realisation xi(s), of the random number of flows, is weighted with the average generated throughput per flow, h(s), for service type s. This results in the total throughput in the backbone for run i ti,backbone = xi(s)T · h(s) [Mb/s] (6.3) The expected value of the backbone utilisation over multiple runs, is then E[ti,backbone] = E[xi(s)T · h(s)] (6.4) = E[xi(s)T] · h(s) (scalar) (6.5) = E[xi(s1)] · h(s1) + E[xi(s2)] · h(s2) + E[xi(s3)] · h(s3). (6.6) This results in three variables to be designed, namely E[xi(s1 = video)] (6.7) E[xi(s2 = voice)] E[xi(s3 = besteffort)]. The random number of flows xi(s) is generated from a uniform distribution with the parameters au,s and bu,s for a target utilisation u for service s. That is, xi(s) ∼ U (au,s, bu,s). (6.8) Hence E[x (s)] = au,s + bu,s