DERIVATION OF THE BANDWIDTH RESOURCE ALLOCATION IN AN ATM- BASED NETWORK

ABSTRACT

Asynchronous Transfer Mode (ATM) technology is the transfer mode for implementing a Broadband-Integrated Services Digital Network (B-ISDN). ATM defines multiplexing and switching  techniques  for  broadband  signals.  Synchronous  Optical  Network  (SONET)  is employed in the B-ISDN backbone. The main objective of an ATM is to guarantee Quality of Service (QoS) in transfer of cell streams across networks. This technology recommended as the transport vehicle for the B-ISDN offers a great flexibility to transmission bandwidth allocation to accommodate diverse demands of multimedia connections. Dynamic Bandwidth Allocation (DBA) is a fundamental factor in network performance for an ATM-based bursty traffic. However, the fundamental problem in ATM network is defining the way to allocate bandwidth optimally  especially  for  unpredictable  bursty  traffic.  This  project  aims  at  developing  the approach to the derivation of the bandwidth resource allocation in an ATM-based network. The main tool used for the simulation is Riverbed Modeler 2014 Academic Edition 17.5 (OPNET). This project developed the intended format with bandwidth allocation guide for N* 64kbps of a Primary Rate Interface (PRI) T1 trunk lines for varying delay, loss and buffer occupancy. The allocation was developed for varying traffic intensity between 0 and 320 Kbps corresponding to five channels of 64kbps each. Cell Loss Probability (CLP) standards between 0.0000875 and

0.002967 were considered.   The Buffer Occupancy values were between 0.00000237KB and

0.0000117KB. Queuing Delay standard which ranges between 0.0000000501s and 0.000000103s and Queue Delay Deviation values of between 0.0000000366s and 0.000000584s were all considered.

CHAPTER ONE INTRODUCTION

In digital communications, bandwidth as a concept has to do with the amount of data a link or network path can deliver per unit of time. For many multimedia applications, the available bandwidth has direct impact on the applications’ performance. The terms bandwidth and throughput often characterize the amount of data that the network can transfer per unit of time [1]. Bandwidth plays a significant influence in several network communications. Several applications can benefit from knowing the bandwidth characteristics of their network paths. Network providers present lists of bandwidth bouquet from which interested users select and are billed. The customers’ subscription to the service provider leads to traffic contract which will finally result in signing of Service Level Agreement (SLA). The rate of bandwidth utilization by various customers makes the providers to plan for capacity upgrade or expansion for the network to avoid congestion, traffic drop or total collapse of the network. It is a standard that bandwidth utilization of above 70% is an invitation to heavy congestion in which case various methods are encourage to avoid such state of congestion. Although network providers can effectively monitor bandwidth utilization through traffic policing and shaping, it is however not the same from the customers point of view. To achieve this network administrator with administrative privileges and access to the network devices such as routers or switches may connect to a link of interest in order to measure the bandwidth using the Simple Network Management Protocol (SNMP). However, such access is typically available only to administrators and not to end users. At times due to congestion which may lead to network failure, end users  can estimate the bandwidth of their links or paths from end-to-end measurements to ascertain the quality of service delivery by the network provider, without any information from network routers due to lack of access. Even network administrators  sometimes  need  to  determine the bandwidth  from  hosts  under their control to hosts outside their infrastructures, which make them to equally rely on end-to-end measurements. There are some bandwidth estimation tools which try to identify the bottlenecks that adversely affect the performance of the network communication.   Some of the publicly available  bandwidth  measurement  tools  include  the  following:  pathchar,  pchar,  nettimer, pathrate, and pathload, AppareNet and lots of other tools. Due to demand by various users, communication network providers, try to allocate bandwidth in order to optimize the network,

enhance network performance and guarantee quality of service delivery to various users whose network demand defer [1].

The above scenario  makes  bandwidth  allocation  a very important  issue in  ATM  networks, especially when there are random fluctuating demands for service and variations in the service rates. In order to make ATM reliable, ATM is designed to support not only a wide range of traffic classes with diverse flow characteristics such as Unspecified bit Rate (UBR) but also to guarantee these traffic classes Quality of Service (QoS) as well. The QoS may be measured in terms of cell loss probability and maximum cell delay [2]. The performance of a network is dependent on the behavior of the QoS parameters. However, the challenge is finding the best way to dynamically allocate bandwidth economically while maintaining low loss and delay [3]. This challenge had necessitated the need to investigate through simulation the QoS parameter of an UBR service category of an ATM network using the Riverbed 17.5 Modeler 2014 version (OPNET). The result of the simulation which was based on Dynamic Bandwidth Allocation (DBA) technique/method was converted into graph using Microsoft Excel. This graph was used to derive the bandwidth resource allocation for UBR service category of an ATM network.

Dynamic Bandwidth Allocation however is a method used in allocating bandwidth dynamically in a network communication technology. It allocates the bandwidth among multiple applications almost  instantaneously  by  providing  each  qualified  service  with  only  its  fair  share  of  the available bandwidth that each application requests at a specific moment. The DBA optimizes the use of available bandwidth without engaging the transmission capacity in advance of which any engagement of bandwidth in advance is static allocation.

One important  feature  of Dynamic Bandwidth  Allocation  is the ability to  make bandwidth changes  based  on  continuous  monitoring  of customer traffic.  If  the  customer increases  the amount of traffic being sent, the algorithm developed should be able to allocate more bandwidth, and if the customer reduces the traffic submitted to the network, the algorithm should reduce the amount of bandwidth that had been allocated to that particular customer [4] and should make such bandwidth available to others in the network.

Network resources such  as the bandwidth  are mainly shared  by users because of distance, population size or human activities (greediness) of not coordinating their actions enough to effectively utilize allocated bandwidth and allow others have their own fair share of the network resources. The reality in communication is that some unreasonable customers have the capacity to  consume  available  network  resources  such  as  the  entire  bandwidth  without  any  sign  of remorse if no constraints were put in place [5]. This consumption of the available network resource affects the QoS performance by introducing congestion and delay to the communication network thereby hindering some more important communication from achieving success. It is instructive to inform that one of the important network resources that requires management in order to avoid congestion and unnecessary delay is the bandwidth. To ensure optimal allocations when  unusual  traffic demands  occur,  any developed  algorithm  needs  to  consider  the time- varying nature of offered traffic which makes such a network a complex dynamic system. In designing such a system, we need to consider the following:

(a) distribution of control functions in networks using Virtual Paths (VPs) (b) monitoring of capacity usage on VPs

(c) calculation of capacity allocated to VPs

(d) frequency with which the capacity can/should be adjusted [6].

From above, dynamic bandwidth allocation is a very serious area which requires appropriate attention and proper management. Bandwidth management is therefore vital in bandwidth allocation.

Bandwidth management can be divided into three categories as follows:

    Bandwidth reservation: Bandwidth reservation dedicates bandwidth to a customer such that even if only a fraction of the reserved bandwidth is utilized, the remaining portion of the reserved bandwidth is not available to any other user in the network.

    Bandwidth limitation: Bandwidth limitation constrains the maximum amount of traffic that can be sent by a single user. If the customer attempts to send more traffic than the

upper limit allows, the traffic could be discarded by the service provider, buffered, or otherwise penalized because it does not conform to the specified limits.

    Bandwidth allocation: Bandwidth allocation is similar to bandwidth reservation explained above because the bandwidth is “guaranteed” to be available to the customer. However, it differs in the sense that if the customer does not use all of the allocated bandwidth, the unused portion is made available to others thereby attempting to dynamically optimising the network. This can be achieved graphically by removing the bandwidth when dormant from redundant user (when not in use) and given to someone that may require it [7].

Bandwidth management can be applied on the customer side or the network side of the ATM User Network Interface (UNI). The network providers are authorized to police the traffic and penalize by discarding and/or delaying nonconforming traffic. To produce conforming traffic, the customer may shape the traffic between the source and the network. Traffic shaping or smoothing refers to queuing ATM cells and then releasing those cells so that the burstiness of the source is controlled.

Although there are many different methods of bandwidth management currently, they are categorised  into  static and  dynamic resource management.  The static  resource management approach does not reflect the inherent changing nature of user requirements which has to do with many variables of which the customer’s state of mind is critical and even the network state which may adjust severally during the lifetime of the connection. This is of great concern to network administrators and company’s management. In addressing the shortcoming above, dynamic bandwidth allocation tends to adjust the bandwidth allocated to a particular customer over time as a result of different variables. This is achieved by shaping the traffic and allocating available bandwidth to other customers in need of it which leads to optimization of the network traffic. The  bandwidth  usage  characteristics  of  customers  can  be  determined  by  monitoring  the submitted traffic.

From  above  therefore,  the  challenge  of  dynamic  bandwidth  allocation  has  gained  lots  of divergent opinions in terms of best practices in its implementation and significant attention in the area of creating new scheduling disciplines and improvement in the areas of existing ones. This means that a lot of work need to be done by the network service providers in ensuring that

various applications from customers are given their fair share of the bandwidth and this can be achieved by appropriate scheduling discipline through available robust scheduling algorithm. On the other hand customers need to use appropriate monitoring tools to ensure that they get a fair share of the bandwidth. The signing of SLA is one approach which can be used in checkmating the behaviour of both sides of the coin. However, putting the content of such SLA contract into reality is another ball-game of which the use of independent bandwidth monitor group whose job can be likened to that of an auditor maybe solicited as a solution with the auditor serving as an arbiter for both sides.

Network   Resource   such   as   bandwidth   is   a   very   vital   and   scarce   commodity   in telecommunication. Due to this singular development, it draws a lot of interest just the way black gold (crude oil) draws in the oil and gas sector. Network operators were interested not only on profit but the Quality of Service rendered to customers especially when they remember sanctions from the regulator when not complying. For corporate organizations requiring bandwidth allocation,  they are  interested  in  quality of  service,  optimal  performance  of  their  allocated bandwidth (best practice for allocated bandwidth) and allocated bandwidth cost implication which is almost the same thinking of single network users. Knowing that BISDN allocates bandwidth in the multiple of 64Kbps (N*64) it becomes a very serious challenge when the actual bandwidth needed by an institution is just 80kpbs (additional 16kbps) or better still 96kbps (one and a half of the normal allocation of 64kbps). If the organisation decides to step down needed bandwidth from 80Kbps to the normal supply of 64kbps bouquet, will the network still perform optimally or will such decision cause network performance issues/failure for the traffic in such organization? Again, if the scale up the bandwidth allocated to 128kbps for either the 80kbps or the 96kbps what will happen to the excess bandwidth of 48kbps and 32kbps respectively? This is a huge wastage for communications’ managers, network administrators and top echelons of organizations and research fellows.

1.1       PROBLEM STATEMENT

Quality of Service (QoS) in an ATM-based network for years has become an important research topic. However, one of the critical challenges in achieving standard QoS requirements is finding acceptable required in dynamically allocating bandwidth resource to various unpredictable demanding multimedia applications.

Therefore within the context of this research, the challenge/problem of this work is to derive bandwidth resource allocation for multimedia traffic for UBR service category of an ATM-based network.

1.2       OBJECTIVES OF THE PROJECT

The objectives of this research work are as follows:

i.      improving network performance,

ii.      providing to the network vendors bandwidth allocation,

iii.       providing  the  capability  to  predict  bandwidth  allocation  and  thereby  improve  trunk utilization,

iv.      contributing to telecommunication knowledge base and forming basis for further research work in the field QoS parameter in ATM network.

1.3       SCOPE

This project was narrowed to modelling, simulation and analysis of QoS parameters for UBR service category in an ATM Network using Riverbed Modeler. The physical medium used for the Network transmission was the topology which is a Synchronous Optical Network (SONET). Simulation result and result analysis, made use of the Riverbed Modeler 2014 Academic Edition

17.5 and Microsoft Excel respectively.

1.4       METHODOLOGY

The methods adopted during the course of the work included studying past literatures in the related  project  field,  developing  both  the  physical  and  simulation  models  for  the  ATM technology, exporting the simulated result to Microsoft Excel for data computation and analysis. Primary research problem focused on monitoring the QoS parameters and traffic (in and out) pattern on each of the switches for the trunk interface Ports (Port 0 and Port 1). Within this perspective, the immediate interest in this platform was ensuring that the satisfied the following network QoS requirements by:

    Providing  bandwidth  guarantees  and  attempting  bandwidth  optimization  which  is  a challenge for UBR;

        Maintaining fairness among the entire network data service;

        Accomplishing increase in the network bandwidth utilization

In  addressing  the  primary  research  problem  in  a  step-by-step  manner,  a  two  dimensional approach was adopted:

    Exploring  the  possibility  of  using  existing  scheduling  algorithm  (Leaky  Bucket)  as informed from the literature review by developing a physical model;

        Transforming the physical model into a simulation model and investigating the QoS

parameter pattern of the Network.

The essence of the simulation model was to underscore the following:

i.      Developing network topology model in Riverbed Modeler 2014 Academic Edition 17.5 and investigating the behaviour of the QoS parameters of UBR ATM network.

ii.      Analysing the simulation result by exporting the graph from the simulation to Microsoft

Excel.

iii.       Studying the behaviour of the trunk paths for each of the switches in other to know to what extent they affect network performance and network utilization.

iv.      Comparing the simulation result gotten from the QoS parameters of the network to the

ATM Forum Traffic Management 4.1 standard.

1.5       THESIS OUTLINE

This work was organised and presented as follows: Chapter two was a detailed literature review on ATM and other related works on ATM. Chapter three discussed the modeling approach adopted and the research methodology used for the work. In chapter four, simulation results of the research work were presented and the result analysis discussed. Conclusion, challenges and recommendations were presented in chapter five.

₦5,000.00 – DOWNLOAD FULL PROJECT DOWNLOAD FULL PROJECT Added to cart

---

---

---

---

---

Related Project Topics :