A fluid queue modulated by two independent birth–death processes
projectsofme Active In SP Posts: 1,124 Joined: Jun 2010 
24112010, 04:49 PM
By: Viswanathan Arunachalama Vandana Guptab S. Dharmarajab A fluid queue modulated by two independent birth–death processes ABSTRACT We present a fluid queue model driven by two independent finite state birth–death processes with the objective to study the buffer occupancy distribution in any intermediate node in a communication network. In a communication network, at any node, the arrival and service of the packets are with variable rates. To model this scenario we develop a fluid queue with an infinite capacity buffer which receives fluid at variable rate and also releases fluid at variable rates. Because of variable inflow and outflow rates of the fluid, the proposed fluid queue is driven by the current states of two independent finite state birth–death processes evolving in the background which on merging give rise to a continuous time Markov chain which is not a birth–death process. Using the fluid queue model, we obtain the steadystate distribution of the buffer occupancy at any intermediate node during packet transmission in a communication network. As a special case, we consider a wireless network based on the IEEE 802.11 standard. We present the buffer occupancy distribution at any intermediate node in closed form with a numerical illustration. Along with buffer occupancy distribution, we also obtain various performance measures such as expected buffer content, average throughput, server utilization and mean delay which are relevant to packet transmission in such a communication network. Finally, we present numerical results to illustrate the feasibility of the proposed model. The results are in accordance with the expected behavior of these performance measures . Introduction In communication networks, information (in the form of data packets) generated by a source node are delivered to their destination by routing them via a multiplexer, a switch, an information processor, or in general, a sequence of intermediate nodes. The information arriving at the intermediate node is buffered for service (transmission), the server typically being a communication channel or processing unit. In highspeed networks, the traffic is very bursty in nature. This bursty nature of the traffic in highspeed networks requires an understanding of steadystate behavior of the system. The steadystate analysis of the buffer content is useful in studying congestion in highspeed networks [1]. In this paper, we use a fluid queue modeling approach to study the buffer occupancy distribution in highspeed networks. Fluid models are a natural choice for problems involving continuous flow. For certain queueing systems where the flow consists of discrete entities, and the behavior of individuals is not important to identify the performance analysis, fluid queue models are useful as approximate models. In high capacity communication networks, the concept of fluid is based on the assumption that most important dynamics depend not on how individual packets are processed, but rather on how aggregates of packets are processed [2]. The applicability of these ideas is based on the fact that the packet size is a very small fraction of typical buffer capacities in the network. Hence, this modeling approach of fluid queues treats the real information flow as a continuous stream rather than considering its discrete nature. Typically, the fluid represents the information stored in a buffer and waiting for transmission in a network. In the fluid queue models, the arrival and service processes are modulated by a random external environment, and the object of interest is to study the behavior of the buffer level in the long run. Fluid queues have been widely used in the performance evaluation of highspeed communication networks [3]. Most of the classical research on stochastic fluid models in the area of telecommunications is based on the work of Anick, Mitra and Sondhi [4,5]. Stochastic fluid models for queues have been extensively studied in [6]. van Doorn and Scheinhardt have analyzed the steadystate behavior of fluid queues which are driven by an infinitestate birth–death process (BDP) [7,8]. Guillemin and Fabrice have discussed the stationary distribution of a fluid queue driven by a finite state Markov chain [9]. In [10], Lenin and Parthasarathy have studied numerically the behavior of fluid buffer driven by truncated BDP with general birth and death rates. A majority of the works aimed in obtaining the buffer occupancy distribution have considered Markov modulated fluid queues wherein the rate of information arriving and leaving the switching component is modulated according to the current state of an underlying Markov process. In [11–13], the authors have used the analogous nature of Quasi BDPs with fluid queues, and the matrix analytic approach of Quasi BDPs to obtain the steadystate distribution of the buffer level. With the fluid queue approach for communication networks, the actual flow of information in a large number of small data packets is modeled as fluid. Following the same approach, we model the flow of information from one node to another via any intermediate node in a network. The objective is to obtain the steadystate distribution of the buffer content at any intermediate node which can give important information on the congestion in the network. The stored information at any intermediate node forms the fluid buffer. The information that arrives at any intermediate node has randomly varying arrival rates which depends on the feedback from other intermediate nodes. The service rates (packet transmission rates) are dependent on the transmission rates of the communication channel. Hence, to model this scenario, we consider a fluid queue which is driven by two different finite state BDPs. We consider an infinite capacity buffer in which the inflow is determined by one BDP and the outflow is determined by another BDP. Note that the two BDPs are independent of each other. We then merge the two background BDPs to form a continuous time Markov chain (CTMC). As a consequence, the considered fluid queue is driven by a single background CTMC (which is a not a BDP). This is a step ahead of the existing literature as in most of the literature on fluid queues, the net inflow rate of fluid into the buffer is determined by a single background BDP [8,10,14–18]. Fluid queue models of this type find applications in communication networks based on the IEEE 802.11 standard. In this paper, we obtain the steadystate distribution of the buffer content at any intermediate node in such a network. In addition to this, we also obtain performance measures like expected buffer content, average throughput, server utilization and mean delay relevant to any communication network. The rest of the paper is organized as follows. Section 2 gives a description of the fluid model. Section 3 presents the application based analysis of the fluid queue model giving the steadystate distribution of the buffer occupancy and various performance measures. Section 4 gives the numerical illustration for the proposed model. Finally, Section 5 concludes the paper with some observations. 2. Model description We consider a Markov modulated fluid queue with infinite buffer capacity. We assume that the buffer is building up and getting depleted with variable rates. To model the variable rate of inflow into the buffer, the inflow rate is determined by a BDP { ˜ X(t), t ≥ 0} with finite state space {1, 2, . . . , N}. Let ˜λi, i = 1, 2, . . . , N − 1 be the birth rates and μ˜ i, i = 2, 3, . . . , N be the death rates of this BDP. When ˜ X(t) is in some state i, i ∈ {1, 2, . . . , N}, then the inflow rate into the fluid buffer is given by ˜ci, which can take any real value. When the buffer level reaches zero and the inflow rate at that time is negative, then the buffer level remains at zero until the inflow rate becomes positive. To model the variable rate of outflow from the buffer, the outflow rate is determined by the states of another independent BDP { ˜ Y(t), t ≥ 0} with M states, 1, 2, 3, . . . ,M. Let ˜αi, i = 1, 2, . . . ,M − 1 be the birth rates and ˜βi, i = 2, 3, . . . ,M be the death rates of this BDP. When ˜ Y(t) is in some state i, i ∈ {1, 2, 3, . . . ,M}, then the outflow rate from the fluid buffer is given by ˜hi, i = {1, 2, . . . ,M}. By combining the above mentioned independent BDPs, we obtain a CTMC with finite number of states. We denote this CTMC by { ˜Z(t), t ≥ 0} with state space S = {(1, 1), (1, 2), . . . , (1,M), (2, 1), (2, 2), . . . , (2,M), . . . , (N, 1), (N, 2), . . . , (N,M)}. This CTMC has NM states. Hence, we have a fluid queue driven by a CTMC (which is a not a BDP). In the next section, we present the analysis of this fluid queue in the context of a practical application. For more information about this topic,please follow the link: web.iitd.ac~dharmar/paper/cma2010.pdf 



Important Note..!
If you are not satisfied with above reply ,..PleaseASK HERE
So that we will collect data for you and will made reply to the request....OR try below "QUICK REPLY" box to add a reply to this pagePossibly Related Threads...  
Thread  Author  Replies  Views  Last Post  
TWO PHASE simplex METHOD PPT  study tips  0  520 
18072013, 04:53 PM Last Post: study tips 

Gaussian Random Processes  study tips  0  297 
20052013, 03:14 PM Last Post: study tips 

One plus one are two  study tips  0  244 
09022013, 11:52 AM Last Post: study tips 

Introduction to Stochastic Processes pdf  project girl  0  406 
21012013, 12:03 PM Last Post: project girl 

Sums of Independent Random Variables  seminar flower  0  434 
24082012, 02:02 PM Last Post: seminar flower 

Twophase flow pressure drop  seminar flower  0  466 
09082012, 03:46 PM Last Post: seminar flower 