Impact Of Burstiness Research Articles

EXPLORING THE IMPACT OF BURSTINESS ON THE SERVICE PROCESS AT THE CASH REGISTER

The service process is the key phase in any queue system applied to business and industry operations. The service process in shops is defined as the payment process at the cash register. The service process consists of two elements or sub-processes: the waiting in the queue to the cash register as well as the payment processing (scanning the goods, giving receipts to customers, etc). Analysis of burstiness as the indicator of the service process has been well-established. Against this background on burstiness as the indicator of the service process, burstiness is also defined as a factor that influences the service process. However, burstiness as a factor in the service process has not attracted a lot of research attention. The aim of this paper is to analyse burstiness as a factor in the service process underpinning the elaboration of scenarios of the service process for the queue management purposes. The present work mostly employs theoretical methods: scientific literature analysis, synthesis, modelling, comparison, and systematization. The theoretical research results in the outline of the conceptual framework for exploring the impact of burstiness on the service process. The key concepts have been identified, namely binary customer behaviour, buyers’ burstiness, bottlenecks’ emergence at the server, and queue management. The logical chain of the development has been emphasized: binary customer behaviour → buyers’ burstiness → bottlenecks’ emergence at the check-out station or cash register (server) → queue management. The presented logical chain allows finding out that buyers’ burstiness leads to the queue appearance in the service process. In turn, queue appearance requires queue management measures. Hence, buyers’ burstiness influences on the decisions in regard to queue management within the service process. Further on, two functions of buyers’ burstiness are defined: the indicators of the service process, and the factor that influences the service process. This bi-modal role of buyers’ burstiness in the service process highlights the complex nature of the queue management. Five scenarios of the service process will allow using a combination of queue management measures in each scenario or even between scenarios. The findings of the comparative study propose the structure of the service process as the unity of the waiting in the queue to the cash register and the payment processing at the cash register, i.e. scanning of the goods and the payment. The present research has some limitations. Further research tends to validate the model of five scenarios of the service process for the queue management purposes. Comparative studies on buyers’ burstiness in the service process will be continued, too.

SSNet: Learning Mid-Level Image Representation Using Salient Superpixel Network

In the standard bag-of-visual-words (BoVW) model, the burstiness problem of features and the ignorance of high-order information often weakens the discriminative power of image representation. To tackle them, we present a novel framework, named the Salient Superpixel Network, to learn the mid-level image representation. For reducing the impact of burstiness occurred in the background region, we use the salient regions instead of the whole image to extract local features, and a fast saliency detection algorithm based on the Gestalt grouping principle is proposed to generate image saliency maps. In order to introduce the high-order information, we propose a weighted second-order pooling (WSOP) method, which is capable of exploiting the high-order information and further alleviating the impact of burstiness in the foreground region. Then, we conduct experiments on six image classification benchmark datasets, and the results demonstrate the effectiveness of the proposed framework with either the handcrafted or the off-the-shelf CNN features.

Base-Station Sleeping Control and Power Matching for Energy–Delay Tradeoffs With Bursty Traffic

In this paper, we study sleeping control (SC) and power matching (PM) for a single cell in cellular networks with bursty traffic. The base station (BS) sleeps whenever the system is empty and wakes up when N users are assembled during the sleep period. The service capacity of the BS in the active mode is controlled by adjusting its transmit power. The total power consumption and average delay are analyzed, and based on this, the impact of parameter N and transmit power on the energy-delay tradeoff is studied. It is shown that, given the average traffic load, the more bursty the traffic is, the less total power consumed, although the delay performance of more bursty traffic is better only under certain circumstances. The optimal energy-delay tradeoff is then obtained through joint SC and PM optimization. The relationship between the optimal control parameters and the asymptotic performance are also provided. Moreover, the influence of the traffic autocorrelation is explored, which shows less impact on the system performance compared with that of the burstiness. Numerical results show the energy saving gain of the joint SC and PM scheme, as well as the impact of burstiness on the optimal energy-delay tradeoff.

The impact of bursty traffic on FPCF packet switch performance

This paper analyses and compares the performance of forward planning conflict-free (FPCF), virtual output queuing-partitioned (VOQ-P) and virtual output queuing-shared (VOQ-S) packet switches. The influence of packet burst size, offered switch load and packet switch size are all investigated. Packet loss probability, average packet delay and 99.9% packet delay values are also derived. In comparison tests, FPCF packet switches perform better than VOQ-P packet switches and yield similar results to VOQ-S packet switches. The advantage of VOQ-S switches is manifested in somewhat lower probability of packet loss, while the advantage of FPCF switches is revealed in the limited values of their maximum packet delays.

Impact of bursty traffic on queues

The impact of bursty traffic on queues is investigated in this paper. We consider a discrete-time single server queue with an infinite storage room, that releases customers at the constant rate of c customers/slot. The queue is fed by an M/G/∞ process. The M/G/∞ process can be seen as a process resulting from the superposition of infinitely many ‘sessions’: sessions become active according to a Poisson process; a station stays active for a random time, with probability distribution G, after which it becomes inactive. The number of customers entering the queue in the time-interval [t, t + 1) is then defined as the number of active sessions at time t (t = 0,1, ...) or, equivalently, as the number of busy servers at time t in an M/G/∞ queue, thereby explaining the terminology. The M/G/∞ process enjoys several attractive features: First, it can display various forms of dependencies, the extent of which being governed by the service time distribution G. The heavier the tail of G, the more bursty the M/G/∞ process. Second, this process arises naturally in teletraffic as the limiting case for the aggregation of on/off sources [27]. Third, it has been shown to be a good model for various types of network traffic, including telnet/ftp connections [37] and variable-bit-rate (VBR) video traffic [24]. Last but not least, it is amenable to queueing analysis due to its very strong structural properties. In this paper, we compute an asymptotic lower bound for the tail distribution of the queue length. This bound suggests that the queueing delays will dramatically increase as the burstiness of the M/G/∞ input process increases. More specifically, if the tail of G is heavy, implying a bursty input process, then the tail of the queue length will also be heavy. This result is in sharp contrast with the exponential decay rate of the tail distribution of the queue length in presence of ‘non-bursty’ traffic (e.g. Poisson-like traffic).

Effectiveness of gain control in EDFAs against traffic with different levels of bursty behaviour

Previous publications have addressed the impact of bursty, self-similar traffic on erbium-doped fibre amplifiers (EDFA). Several gain control techniques have been suggested to combat the gain fluctuations of EDFA with this type of traffic. The effectiveness of two such methods are investigated, namely highly inverted amplifiers and all-optical gain clamping, while varying the parameters characterising the burstiness of the sources. While previous publications have focused on the effect of the average load or activity factor, the paper further investigate the dependence on the relative variability of the packet burst lengths (ON times), and lengths of the interburst idle times (OFF times). Both gain control methods reduce the output power and signal-to-noise ratio excursions with respect to a standard amplifier chain. The authors find that, for an eight channel WDM system and a cascade of six EDFAs, all-optical gain clamping can reduce the variations by a factor of five, provided adequate power is present in the clamping laser, while highly inverted amplifiers have variations reduced by a factor of two. The authors find that, in a clamped chain, the source activity factor does not uniquely determine the required level of lasing power. The authors also deduce that information on the relative variability of ON and OFF times is essential.

Performance analysis of an ATM statistical multiplexer with batch arrivals

The authors propose a method for evaluating the impact of bursty traffic sources on the performance of an ATM multiplexer which can handle packetised voice, data and video traffic. The multiplexer is modelled as a single-server queue with batch phase type Markov renewal input. The authors describe an analytical approach to the solutions of such queueing systems with finite capacity using the matrix analytic method. The methodology was chosen because its ability to model exactly complex nonMarkovian models having correlated, nonrenewal input processes. It is nearly impossible to obtain exact expressions using the classical method of generating functions. A study of different mixed traffic assumptions indicates that bursty traffic does have a very significant effect on the loss performance.