Random multiple access of broadcast channels with Pareto distributed packet interarrival times

Abstract

We consider the random multiple access of a slotted broadcast communication channel. Packet arrivals for such channels are often modeled as Poisson processes because the latter have attractive theoretical properties and are well understood, even though a number of tra c studies have shown that packet interarrival times are not always exponentially distributed. For example, recent studies argue convincingly that tra c from a variety of working packet networks (e.g., LANs, WANs, e.t.c), is much better modeled using statistically self similar processes to characterize the packet interarrival times of the aggregate tra c, which have much di erent properties than Poisson processes. Because of the great in uence the multiple access algorithms (used in the channel access subsystem) will have on the performance of third generation digital wireless mobile communication systems, we study in this paper the problem of random multiple access under a packet tra c generated by interactive data applications (e.g., reading/composing short email messages, responding to paging type messages, transferring les between the mobile and a xed computer, and querying of a database). We examine and demonstrate the performance robustness of the above RAAs when they are driven by a strongly bursty packet arrival process, commonly found in interactive data applications.

Analysis of a random-access protocol under long-range-dependent traffic

“…In this paper, we show that when the LRD traffic is not very bursty, the performance of the random-access channel can be much better than that produced under Poisson traffic loading; otherwise, a distinct performance degradation takes place. In light of these results, we conclude that neither the LRD Telnet arrival processes used by Aracil and Munoz [13] nor the Pareto traffic processes used by Harpantidou and Paterakis [14] are very bursty. In fact, by re-examining the model used in the latter paper, we have found that in constructing the Pareto traffic model, Harpantidou and Paterakis have selected the parameters for the Pareto traffic model in such a way that the modeled traffic becomes less and less bursty when the channel loading gets heavier.…”

“…Furthermore, one expects such LRD properties to have a significant impact on the performance of Aloha-based channels as well. This issue has been recently studied by Aracil and Munoz [13] and by Harpantidou and Paterakis [14]. By employing LRD real Telnet packet-arrival processes as input traffic, Aracil and Munoz have found that the Aloha channel actually performs much better than that loaded by a Poisson traffic.…”

Visualization of Mobile Network Simulations

“…Therefore, multi-layer adaptive techniques are being pursued to coordinate the resource management task among protocols at multiple layers within the communications architecture [7,8]. Finally, due to the complexities inherent in modeling a heterogeneous mobile network, simulation has become a primary method for performance analysis of mobile network protocols [4,5,8,9,10,11,12].…”

Medium access control and traffic policing for multimedia integrated access in high capacity wireless channels

“…1) The aggregate data message arrivals are Poisson distributed with mean λ messages per frame, and 2) the interarrival times of the aggregate data message arrival process are assumed independent, and identically distributed according to a Pareto distribution with shape parameter α and location parameter k. We use the bursty Pareto model because it simulates well data traffic generated by interactive applications, the importance of which is growing and will be even more significant in future mobile systems [20,21]. The location parameter k corresponds to the minimum interarrival time between packets.…”