摘要 ABSTRACT:

預計未來一代的無線網絡將支持多種無線接入技術,每個都有不同的訪問帶寬和覆蓋范圍。這些技術包括通用移動通信系統(UMTS)和IEEE 802.11無線局域網絡(WLANs)。為移動用戶提供實時無處不在的溝通服務,隨時隨地使用這些異構網絡，目前看來仍然是一個挑戰。
 

這項工作提出了一個端到端的移動性管理方案,使移動用戶維持無縫連接,同時越過這樣的異構接入網絡。持續的網絡連接是通過使用配備雙網絡接口的移動設備獲得的，這些接口之間的數據傳輸的能力取決于訪問網絡的可用性。我們提出并設計了這樣一個無線接入設備的雙核模式，提出了一個簡單的交叉切換技術。我們使用OPNET仿真測試評估我們體系結構的性能。
 

1.前言

移動通信和無線網絡的發展速度迅速。在這兩個學科范圍內，先進技術層出不窮。迫切地需要建立一個強大的集成無線局域網，開發混合動力移動數據網絡，提高熱點數據服務，保證高數據率。
 

Future generation wireless networks are expected to support multiple wireless access Technologies, each with different access bandwidth and coverage range. Two of these technologies include Universal Mobile Telecommunications System (UMTS) and IEEE 802.11 Wireless Local Area Networks (WLANs). Provision of real time ubiquitous communication to mobile users, anywhere, anytime through these heterogeneous networks remains a challenge.
 

This work proposes an end-to-end mobility management solution that enables mobile users to maintain seamless connectivity while moving across such heterogeneous access networks. The continuous network connectivity is obtained through the use of mobile devices equipped with dual network interfaces and the capability to switch data transmission between these interfaces depending upon the availability of the access network. We present and discuss the design of such a dual- mode radio access device and propose a simple inter-technology handoff technique. We evaluate the performance of our proposed architecture by conducting experimental simulation tests using OPNET.
 

1. INTRODUCTION

The Mobile communications and wireless networks are developing at a rapid pace. Advanced techniques are emerging in both these disciplines. There exists a strong need for integrating WLANs with UMTS to develop hybrid mobile data networks capable of data services and very high data rates in hotspots.
 

UMTS systems such as Universal Mobile Telecommunication Systems (UMTS) can provide mobility over a large coverage area, but with relatively low speeds of the one hundred and forty four kilo bit per second. On other hand, WLANs provide high speed data services (up to 11 Mbits/sec with 802.11b) over a geographically smaller area. WLANs are generally used to supplement the available bandwidth and capacity of a UMTS network in hotspot areas such as railways and airports with high traffic-densities, without sacrificing the capacity provided to cellular users.#p#分頁標題#e#
 

The rest of this project is organized as follows. Section 2 provides a brief background on UMTS and WLAN networks. Section 3 provides the integrating architecture of WLAN and UMTS. Section 4 provides us the methodology and the design model of the simulation. Section 5 provides us the analysis and results of the simulation. Section 6 discusses the design and simulation of the proposed switching and handoff techniques for roaming between UMTS and WLAN networks. Section 7 we compare two integration architectures connecting UMTS and 802.11 networks.
 

2. BACKGROUND

2.1 Standards of WLAN

These standards WLAN [1] have been deployed in offices, homes and public hotspots such as airports and hotels given its low cost, reasonable bandwidth (11Mbits/s), and ease of deployment. However, a serious disadvantage of 802.11 is the small coverage area (up to 300 meters) [2]. Other 802.11 standards include 802.11a and 802.11g, which allow bit rates of up to 54 Mbits/sec.
 

2.2 UMTS Networks:

ITU defines UMTS as any device that can transmit or receive data at 144 Kbps or better [3]. In practice, UMTS devices can transfer data at up to 384 Kbps. As a comparison, Global System for Mobile Communications (GSM) data rates are up to 14.4 Kbps and General Packet Radio Service (GPRS) is around 53.6 Kbps used in 2G and 2.5G respectively.
 

Two main proposed systems for UMTS recognized by the International Telecommunication Union (ITU) are Code Division Multiple Access (CDMA 2000) and Universal Mobile Telecommunication System (UMTS) [4]. UMTS is composed of two different but related Modes. Wideband CDMA. FDD mode is considered the main technology for UMTS. Separate 5 MHz carrier frequencies are used for the uplink and downlink respectively, allowing an end-user data rate up to 384 Kbits/sec.
 

2.3 UMTS Network Architecture:

A UMTS network consists of three interacting domains- a Core Network (CN), Radio Access Network (RAN) and the User Equipment (UE) also called a UMTS mobile station [3, 13]. UMTS operation utilizes two standard suites: UMTS and CDMA2000 which have minor differences with respect to the components they have in the RAN and the CN.
 

The main function of UMTS core network is to provide transit, routing and switching for user traffic. It also contains the databases management functions. The CN is divided into Circuit-switched (CS) and Packet-switched (PS) domains. The elements of Circuit switched include Visitor Location Register (VLR), and gateway MSC. These entities are common to UMTS as well as the CDMA2000 standards. The differences in the CN with respect to the two standards lie in the PS domain. Packet-switched elements in UMTS include (SGSN and GGSN)
 

CDMA2000 packet-switched component is primarily the Packet Data Serving Node (PDSN). Some network elements like Equipment Identity Register (EIR), Home Location Register (HLR) are shared by both domains. The main function of the MSC server is to handle call-control for circuit-based services including bearer services, etc. The MSC server also provides mobility management, connection management and capabilities for mobile multimedia as well as generation of charging information. It can also be co-located with the Visitor Location Register (VLR).#p#分頁標題#e#
 

GGSN is the gateway to external data networks. It supports control signalling towards external IP networks for authentication and IP-address allocation, and mobility within the mobile network. GGSN provides functions for forwarding and handling user information (IP packets) to and from external networks (Internet/intranets). SGSN provides session management, i.e. mechanisms for establishment, maintenance and release of end user Packet Data Protocol (PDP) contexts. It also provides mobility management and supports inter-system handoff between mobile networks. SGSN also supports generation of charging information. The PDSN incorporates numerous functions within one node. Routing packets to the IP network, assignment of dynamic IP-addresses and maintaining point-to-point protocol (PPP) sessions are some of its main functions. It also initiates the authentication, authorization and accounting (AAA) for the mobile station. The radio access network provides the air interface access method for the user equipment. And UMTS RAN (UTRAN) consists of Radio Network Controllers (RNC) and Base Stations (BS) or Node-B. The RNCs manage several concurrent Radio Link Protocol (RLP) sessions with the User Equipments and per-link bandwidth management. It administers the Node-B for congestion control and loading. It also executes admission control and channel code allocation for new radio links to be established by the Node-B.
 

A CDMA2000 RAN consists of a base station and 2 logical components- the Packet Control Function (PCF) and the Radio Resources Control (RRC). The primary function of the PCF is to establish, maintain and terminate connections to the PDSN. The PCF communicates with the RRC to request and manage radio resources in order to relay packets to and from the mobile station. It also collects accounting information and for wards it to the PDSN. RRC supports authentication and authorization of the mobile station for radio access and supports air interface encryption to the mobile station.
 

3. WLAN AND UMTS INTEGRATION ARCHITECTURES

UMTS and WLAN Networks have been integrated with different approaches and strategies. The two most commonly used approaches are tight and loose internetworking. Other strategies are modifications of these two basic approaches. In the tight internetworking approach, the WLAN network does not appear to the UMTS core network as an external packet data network. Instead, it appears simply as another UMTS Radio Access Network. WLAN network in this case emulates several functions of the UMTS RAN. This is made possible by employing a specialized WLAN gateway in between the UMTS core network and the WLAN network that hides the details of the 802.11 network and implements all UMTS protocols required in the UMTS Radio Access Network. The architecture is capable of providing roaming services to users with dual stack (UMTS and 802.11) network cards in their mobile devices. Using this approach, both the networks often share common billing and authentication mechanisms. However, all traffic from the WLAN network passes through the UMTS core network, which could cause it to become a bottleneck.#p#分頁標題#e#
 

Tight internetworking also requires common ownership of the two networks that does not make it a very feasible deployment strategy. However, tight internetworking does offer high security mechanisms as the UMTS security protocols can be reused in the WLAN network. It also provides fast handoffs as roaming between the two networks is the same as moving between two RANs of the same UMTS network (since the WLAN network appears as a different Routing Area only).
 

Loose internetworking, on the other hand the WLAN gateway directly connected to the Internet and does not have any direct link to UMTS network elements. In contrast to tight coupling, the WLAN data traffic does not pass through the UMTS core network but goes directly to the IP network (Internet). In this approach, different mechanisms and protocols can handle authentication, billing and mobility management in the UMTS and WLAN portions of the network. Loose coupling has low investment costs and permits independent deployment and traffic engineering of the WLAN and UMTS networks. Other UMTS-WLAN internetworking strategies and their features are summarized in the given table. The internetworking architecture of Mobile IP [5] considers WLANs and UMTS network as independent. Which allows the easy deployment of but suffers from functions, real tome services and long handoff latency. The gateway approach [10] permits the 2 network independent operation which then gives facility of seamless roaming between the networks. Both networks are connects via virtual GPRS support node using a new logical node. The use of Mobile IP is not required in the gateway approach which has a loss during handoff comparatively lower packet. It is difficult to deploy emulator approach [5] when it requires ownership of the both two networks but due to low handoff latency yield make it much better suited for real-time applications well suited.
 

At last the peer-networks [12] approach allows the deployment easily but also suffers from

Table 1: UMTS-WLAN internetworking strategies Comparison

Internetworking

Approach

Handoff

Latency

Network Management

and Ownership

Authentication

Billing

Real time

application

support

Mobile-IP

approach

High

Separate ownership of

UMTS and WLAN networks

permitted with agreements of roaming

UMTS cellular AAA rescue

functions WLANs

Billing

Feature UMTS users

No

Gateway

approach

Low

Permits UMTS and WLAN

networks to operate

independently

Can provide both, separate

WLAN security as well as

UMTS AAA

Can provide

separate as well as

combined billing

mechanism

Yes

Emulator

approach#p#分頁標題#e#

Low

Combined ownership

required

Can provide both, separate

WLAN security as well as

UMTS AAA

Uses UMTS billing

mechanism

Yes

Tight coupling

Low

Generally requires both

WLAN and UMTS networks

to be owned by same

operator

UMTS ciphering key used

for WLAN encryption

UMTS accounting

features re-used

Yes

Loose

coupling

High

Permits independent

deployment of WLAN

and UMTS networks.

Cellular access gateway

provides authentication

Billing mediator to

provide common

accounting

Not very

Suitable

Peer networks

approach

High

Both, same or different

operator ownership

permitted

The Use of AAA functionality

of the UMTS network

Billing feature of

UMTS network

used

Not very

suitable

high handoff delays, thereby making it unsuitable for real-time applications. The choice of the integration architecture is important since multiple integration points exist with different cost- performance benefits for different scenarios.
 

4. METHODOLOGY AND MODEL DESIGN

4.1: Performance Matrices:

Both the systems are working at the data session set up delay, interaction of protocol along the dedicated channel utilization (DCH), time of download response FTP and web page. The system control level of interaction protocols can plane signalling message by transactions which is between the nodes of the systems. At the model verification process interaction metric was simulation and used between the both network UMTS and WLAN is accurately modelled.
 

The active UMTS number of dedicated channels is known as the DCH utilization c. After the measuring of different channels the small value is consider best and in the UMTS and WLAN system this is proposed. When the first response packet is received between an application and clients requesting service to the time is known as setup delay of data session and consist of PDP context, time to establish a TCP connection the Radio Access Bearer. The elapses time for getting a request between receiving and sending of the file FTP response time and setup connection is also included in it. The web page response time to retrieve the HTML page along with its component in the required time. When application is reduced to Access Network of WLAN to shifting the benefits of data users are demonstrated and were measured by both application delay performance. In the UMTS DCH cell state the UWs and UEs were run in the simulation which shows that downlink and uplink traffic of the UMTS was sent to the dedicated channel of UMTS.#p#分頁標題#e#
 

4.2: SIMULATION FACTORS:

The simulation factors in this integration are application profile, number of clients, access mode, and number of clients while the application profiles
 

Using the Wireless Client Light, Traffic Models, FTP Fixed File Size, Wireless Client Heavy and the simulation was run. For a particular client more then one profile can be cond. From one to hundred MB FTP file size was running the simulation. Due to simulation factors both the UMTS DCH channels and WLAN determine the traffic channel. The UW in UMTS and WLAN mode is used by Interface of WLAN. With thirty clients to gain access the simulation was run to the channel of WLAN and the client access mode setting is depend on it. For both traffic loads can be measured by combining profile of application.
 

4.3: Technique of Evaluation

The outcome can greatly impact the performance evaluation of technique of this selection. Measurement, analytic and simulation are the performance technique of evaluation and in cost, accuracy and required time are not the same as. The less accuracy is given by the solution of analytic and is consuming time and is cheap as compared to the simulation. And due to this simulation the analysis of this performance is avoided. In the model verification process analytic method is used.
 

4.4 Traffic Model

The wireless data user generates traffic used by the OPNET™ application profile to closely simulate. E-mail, FTP and HTTP profiles were used in application profile. File transfer protocol send file in between server and client. The file get and put command allow user for transferring data and for which transfer protocol is TCP by default. Type of service, inter request time file size and the command mix are the attributes of FTP. POP and simple mail transfer protocol of email application use a default TCP.
 

The email application includes the attributes are Type of Service, Receive Group Size, Send Group Size, Send Inter arrival Time and email size.
 

Web browsing is HTTP application and from a server page is downloaded by the users including graphics and texts. Multiple TCP connections are open in result a single HTTP page request. The http includes the attributes of Pages per Server, Server Selection-Initial Repeat Probability, Pipeline Buffer Size, Page Properties-Object Size, Type of Service Max idle period and HTTP Specification.
 

4.5: Simulation Scenarios:

Access network used is the basic difference between the three scenarios. The traffic model and simulation factors varied these scenarios. The performance here is measured by performance metrics.
 

In Existing WLAN vs WLAN-UMTS Scenario the FTP application running by a user a single wire client simulation. The aim of the scenario is the difference interworked WLAN-UMTS system from the existing one. The application response time data session set-up time were evaluated for both systems. The network traffic load varies with size of FTP. A client can access the network by access point in the simulation by the Existing UMTS vs WLAN-UMTS scenario. In term of application response time and DCH utilisation the above systems were evaluated. Through UWLAN_AP number of UWs getting the network of UMTS on the simulation of WLAN-UMTS. Client running FTP application measures the download response time for both.#p#分頁標題#e#
 

4.6: Simulation Design

The OPNET Modeler™ 9.0 was used for the approach of top-down design in this simulation. The hierarchical structure of processes, nodes and network scenarios are used in it. The user defined and built-in OPNET™ processes were used in the design of UW and UWLAN_AP. State diagram uses process level that describe the UW_MAC_IF_Contro and UWLAN_AP Controller. The built-in OPNET functions and C code defines the user defined process actions in every single simulation. To gain network access through UMTS Node B the UW was augmented. An alternative radio access is formed due to the Interworking the WLAN technology into UMTS as created by the UW and UWLAN_AP in the simulation frame work.
 

4.6.1: Network Level

Through UWLAN_AP to CN at the gains access by UW through the RNC to Node B at UMTS client access WLAN client access mode. Initialisation of network is at start up of the simulation. At the authentication of UMTS network when UW power on for GMM_Attach procedure performing. UW is register for the services of GPRS network in GPRS Mobility Management (GMM) Attach procedure by establish a PS signalling connection with the SGSN by notifying SGSN. Explicitly models between UE and SGSN of GMM attach procedure in the OPNET™ UMTS model set.
 

4.6.2: The Enhanced WLAN Access Point Node

This includes the UMTS specific ATM interface module, physical (PHY), WLAN medium access control (MAC), a, and user-defined process module and a full ATM stack. The standard model set of WLAN MAC and PHY gives implementation of the 802.11 standard of protocol level. ATM and AAL5 protocols are implemented by the ATM stack. UMTS specific ATM interface provided by UMTS specialized model set that maps ATM and UMTS QoS and between the components of UMTS controls the virtual circuit.
 

The Interworking of WLAN infrastructure BSS at the RNC is the responsibility of user-defined UWLAN_AP_Control process module and by teardown and RB (Radio Bearer) installation and RAB (Radio Access Bearer) by controlling.
 

4.6.3: UWLAN_AP_Control Process Module:

An alternate radio access network into UMTS system by the by the control logic of WLAN interworked is provided by this module. To build the access control table which perform GMM attach by UW in the Modified_GMM_Attach_Request message in the UWLAN_AP. Using the M-state , Mac address and triplet IMSI adds a UW entry to WLAN_AP. The RAB_Assignment procedures and PDP_Context_Activation are implemented in this process. Four control and three initialization states are in this state diagram.
 

States functions are given below:

INIT: Static's can be collected in the stage variables which are initializes in this stage process in the model-wide process for registry. Wait and init2 schedule self interrupts to initialise ATM PVCs and lower layer WLAN MAC.#p#分頁標題#e#
 

Idle: Packets arrival from the ATM interface and WLAN MAC are associated in this. The packet associated with UWs IMSI and UMTS message type is responsibility of this state.
 

From_UW: The transmission of data by UMTS data traffic is authorised by the UW. The signalling message and data packet processing and Activate_PDP_Request are sending to the ATM interface to the CN.
 

From_CN: The transmission of data by UMTS data traffic is authorised by the UW. The signalling message and data packet processing and Activate_PDP_Request are sending to the WLAN interface the UW.
 

ADM_CNTL: The action required to take by this in the signal messaging handle is by ADM_CNTL. UWLAN_AP monitor track of the UMTS authorized UWs that is build by the access control of the ADM_CTRL and coordinates in the setup procedure of RAB.
 

4.6.4: Enhanced User Equipment Node (UW):

The addition of UE workstation model to WLAN protocol stack for the client access configuration of necessary control logic by UMTS model augments in the model of UW. The UW_MAC_IF_Control and user-defined process module are implemented by control logic.. All are shown in the below .
 

High-fidelity implementation of network and transport layer is provided by standard OPNET™ TCP/UDP/IP protocol stack. The UMTS control plane the UMTS control plane provides the detail implementation of GMM module and works between SGSN and UW and gives authentication to UMTS PS domain of UW and mobility management. Medium access control layer and UMTS Reliable Link Control are implemented and provided by RLC/MAC module. The radio accessing signals are implemented by the MAC portion. The radio interface switching is between the WLAN and UMTS radio stack. The setting of Client Access Mode to WLAN and WLAN MAC to the UW_MAC_IF_Control through the UWLAN_ AP when using WLAN.
 

4.6.5 Core Network (CN):

The core network contains the SGSN, GGSN, visitor location centre, authentication centre and connectivity to RNCs and UWLAN_AP provided by 8 ATM layer interface. The IP routing stack is modelled in the GGSN of IP gateway functionality.
 

4.6.6: SGSN Process Module:

The Session Management (SM) protocols and GPRS Mobility Management (GMM) are the 2 system networks models of this process module. The packet data protocol (PDP) known as packet data session are established due to various Service Request_* and PDP_Activate_* messages at the network of UMTS. Radio Access barer is established between its SGSN and UE.
 

4.7: Simulation Verification and Validation:

The verification and validation of the sections ensure that the simulation is corrected in representative and implementing.
 

4.7.1: Model Verification:

The process which determines that the functions of simulated model are working correctly is known as verification model. Testing for logical errors, computer codes debugging, different modules testing functionality includes in this. Due to the interactive control of the users the ODB (OPNET Simulation Debugger) provides investigation of the simulation through break points to give information of objects and events. For every message trace is created for the verification of access control protocol and user-defined authentication. In the appendix the results of traces can be seen. The verification process is passed through the access control protocol and user defied authentication. The WLAN and UMTS models are tested which is required. For various WLAN and UMTS nodes for the correct operation of verification short simulation were run. In the same way both models were verified. The OPNET™ Technologies model documentation results were compared to the simulation results [25].#p#分頁標題#e#
 

4.7.2: Model Validation:

The theoretical results or real system measurements by using expert intuition can validate the simulation [22]. Real system measurements and results of out put simulations is the best way to model validating.
 

Implementation dependent in backside server network gives delay. A constant delay in slow-start algorithm and TCP connection setup is assumed. The PCF signalling delays and link transmission delays were considered in small delays and considered to zero there node delay.
 

Calculation in matleb was theoretical results and simulation results are close to it.
 

5. Results and Analysis

The result of simulation and their analysis are discussed in this chapter.
 

5.1 Statistical Accuracy:

In simulation the measure of uncertainty is introduce in the stochastic processes. The different generator seeds give different results in the simulation. Every set of input parameters were run at 5 different seeds in the simulation model of this research.
 

The single UW in the WLAN Client results for FTP download in the FTP application traffic is given. For the all file sizes there is a difference of one percent in FTP response time between minimum and maximum.
 

FTP Filesize

1 B

100KB

800KB

1.5 MB

7MB

12MB

15MB

30MB

40MB

60

Seed 1

0.194574

1.177744

6.337744

11.71774

52.45774

89.95774

112.5177

223.9377

298.0577

448.7377

Seed 2

0.188612

1.190725

6.350725

11.73073

52.47073

89.97073

112.5307

223.9307

298.0707

448.7507

Seed 3

0.183707

1.203707

6.363707

11.74371

52.48371

89.78371

112.4086

223.7437

297.8837

448.5637

Seed 4

0.190725

1.068612

6.428612

11.60861

52.54861

89.84861

112.3437

223.8546

297.9546

448.6086

Seed 5

0.197744

1.094574

6.454574

11.63457

52.57457

89.87457

112.4346

223.8086

297.9486

448.6346

Mean

0.191072

1.147072

6.387072

11.68707

52.50707

89.88707

112.4471

223.8551

297.9831

448.6591

95%CI:

0.184352

1.071135

6.323218

11.61113

52.44322

89.79029

#p#分頁標題#e#

112.3503

223.7529

297.8847

448.5574

0.197793

1.22301.

6.450927

11.76301

52.57093

89.98386

112.5439

223.9573

298.0815

448.7607

Table 2: WLAN Client Access and FTP Download Response Time

The next one is for the WLAN-UMTS scenario that gives HTTP page response.

UWs

20

18

16

14

12

10

08

06

04

02

Seed1

12.432839

9.167257

8.251739

8.116110

8.385436

8.096770

7.359587

7.627617

8.387927

7.525191

Seed2

13.416985

8.911012

8.658675

8.086360

7.534312

8.096770

7.651648

7.722649

7.751405

8.666523

Seed3

13.667978

8.865238

9.02389

8.796989

8.571726

7.673093

7.694435

8.272729

7.446080

7.268593

Seed4

12.710351

9.249344

8.772352

10.255774

8.036825

8.029945

7.695623

7.315788

7.261536

7.848843

Seed5

13.161796

8.891231

8.517017

8.068062

8.243471

7.538070

7.461991

7.967559

7.586461

7.821761

Mean:

13.077990

9.016817

8.644735

8.664659

8.154354

7.886930

7.572657

7.781268

7.686682

7.826182

95% CI:

12.450649

8.795917

8.287612

7.496688

7.660177

7.560854

7.382510

7.333671

7.151045

7.172729

13.705330

9.237717

9.001858

9.832630

8.648531

8.213006

7.762803

8.228865

8.222319

8.479636

Table 3: UMTS Client Access Mode and HTTP Response Time for UWs

With the different seed value the mean value of these simulations runs with the all data points in this section. In appendix the complete individual simulation are present.
 

5.2 THE Existing WLAN-UMTS Vs UMTS Scenario Results:

The UMTS system is compared to the benefits of the Existing one with the scenario WLAN UMTS is demonstrated in this section.
 

5.2.1 Dedicated Channel (DCH) Utilization:

To continue providing services the main aim of this integrating was the resources to free them that's why the node B was examining for the for dedicated channels numbers due to the network of UMTS finite resources and importance of voice channels. With the client data of 20 wireless users this test was conducted with the Email and FTP applications and specifically the mix of web. Operation in the cell of UMTS the 20 UW nodes are present in the baseline system. For all 20 clients UMTS was set for the first run simulation. Approximately sixty five percent of the DCH channels are reduced which can be seen in the below.#p#分頁標題#e#
 

The two different traffic profiles results are shown in the given subsections:
 

5.2.2 Application Response Time:

In WLAN Client Access Mode UW operating was replaced by UE node. The available bandwidth was increased for the hot spots data user in this WLAN and UMTS integration. The decrease in the response time of FTP download can be presented in the below diagram.
 

As the dedicated channels are not allocated dynamically that's why the available UMTS channel was used insufficient by the system. The DCH channels are allocated by the RNC dynamically.
 

Wireless Client Light Traffic Profile

This application shows that the web page response and the FTP download there is a significant decrease WLAN access network in the UMTS data services. Systematically the number of users of wireless data users varies in both the system scalability. The conduction of the test was in Wireless_Client_Light application profile with no of data users.
 

The above shows that 2 to 30 users were increased in the wireless data users and in the same graph the response of both were plotted as well. When the users were more then twenty seven then the performance of FTP is little decrease which can be seen in the .
 

When the users were greater than 25 then there is a sharp increase in the UMTS access network in application response time because the network saturation experience starts with 18 users. A significant increase is in queued request for services when the users are increased to 30.
 

5.3 Results Existing WLAN vs WLAN-UMTS

The UMTS signalling were incorporated as result of delay in the initial setup in the encounter of UW and the performance measure of the initial setup delay are proposed in the system of WLAN and UMTS per packet.
 

5.3.2 Delay of Application

The WLAN infrastructure BSS with UMTS session management protocol interworked additional per packet delay is demonstrated in this section. The network traffic load increases systematically with FTP file size. The node of WLAN was replaced by node of UW through WLAN network access in UMTS utilizing.
 

The network access is gained by the WLAN client through WLAN Access Point to download 30 Mega byte file in sixty five seconds.
 

5.4 UMTS Results Mixed Client Access WLAN

In normal range of operating conditions in the interworked system of WLAN - UMTS of the above mentioned scenario. The UMTS signalling protocol was demonstrated correctly in this section.
 

5.4.1 UMTS Signalling:

In UMTS and WLAN client access mode attach GMM service activation delays and PDP context activation was demonstrated in this section. With in 300 meter twenty clients were connected to the WLAN access network and operating through access point of WLAN.#p#分頁標題#e#
 

UMTS PDP Context Signalling

The UMTS PDP context activation process operation was demonstrated in this section. The class traffic and requested quality of services profile are included in the PDP context activation. RAB establishment for the active PDP was used by Service Request procedure.
 

When the resources are available due to the serviced higher priority traffic the lower priority has to wait and established sixty PDP contexts in the configuration.
 

UMTS Service Request Signalling

The UMTS service request operation was demonstrated correctly in this section. In QoS the establishment of PDP context couldn't be deactivated by the model and reused the PDP context for data sequence subsequent. The procedure of service request used in RAB establishment associated with QoS upon data packet receipt.
 

5.4.2 The Traffic Load Wireless Client Light

The UMTS access network and UWLAN-AP accessed the services of UMTS data in UWs. UMTS parameterized traffic load capabilities are interworked system of WLAN and UMTS demonstration is the purpose of this section. In simulation client access vary from zero WLAN and twenty UMTS, twenty WLAN and zero UMTS, eighteen WLAN and 2 UMTS. The UMTS access networks and UWLAN_AP varying the network load results systematically.
 

For services there were 0 queued for UW assessing the UMTS network but the UWs number were increased and dramatically increased in Queued request service and the denied or delay increased for the requested service as number increases.
 

The FTP application traffic results similar demonstration in the given . Accessing UMTS network takes large for the accessing UWs in FTP download response time. Half were accessing the WLAN access network at instance operating point and half were accessing UMTS and download web page in sixteen seconds.
 

5.4.3 Wireless Client Heavy Traffic Load:

Traffic load were heavier significantly in WLAN access network capabilities which are demonstrated in this section. In order to stress the WLAN channel file of size and with greater object with higher rate of email request was parameterized in wireless client application profile. The number of UWs varies from two to thirty for FTP download response time as shown in the below .

The application response time has significant impact beyond the adding user over twenty three running mix application traffic. Twenty users operating at thirty seven seconds in FTP download response time.
 

Downloading a web page with 2 seconds average per user of the twenty users and in sixteen seconds when users are increases to thirty
 

When ranging the two users from two users of WLAN offer load time is varying which can be seen in the given . WLAN load increases with 1.2Mb/sec with no load and twice at peak to 4 MB/sec with two users. The needed bandwidth were getting by the users of lightly loaded by the WLAN access network and this result was produced by the twenty and twenty three users but reached 7 MB/sec at peak and to reach saturation at WLAN channel caused by thirty users. In FCP mode maximum through put approach to 11Mb/sec.#p#分頁標題#e#
 

6. DESIGN OF DUAL-MODE TERMINAL AND HANDOFF MECHANISM

A network simulation model was constructed using OPNET 10.0.A simulator [10]. Fig. 4 illustrates the network level view of the simulated network topology. It is composed of a WLAN network (hotspot) of radius 150 meters located within a UMTS cell. The two networks are integrated at the GGSN node. The topology chosen for our simulation tests corresponds to real-world scenarios where a WLAN network serves as a hotspot within a UMTS network. Such hotspots can be found in office buildings, hotels, airports, train/bus stations, cafés etc. WLAN is a low cost broadband technology; however it does not provide end-to-end coverage. The topology can be a good example of a scenario where a corporate user uses the WLAN in his office building and as he steps out of the building, he uses the UMTS network (in this case UMTS). The WLAN network is within the coverage of a UMTS network. Another example that reflects such a setup is as follows: a typical user on the way to his office might roam from a home WLAN network onto a UMTS network then onto a public WLAN hotspot, say in a cafe, and then back onto a UMTS network and then finally onto an enterprise WLAN. Thus the description of the 3G-WLAN architecture simulated topology closely represents actual user mobility scenarios. The different entities of the two networks (WLAN and 3G) described in the topology are however, generic to all cases. The UMTS network is composed of a UTRAN (RNC and Node B) and a UMTS core network (SGSN and GGSN). Details of 3G network components are outlined in [3].
 

RNC: Radio Network Controller SGSN: Serving GPRS Support Node

GGSN: Gateway GPRS Support Node WLAN: Serving GPRS Support Node.
 

In the simulated network topology, the mobile node is a Dual Mode Terminal (DMT) with both a UMTS interface as well as a WLAN interface. We have designed and implemented a software layer called the Switching Module incorporated into the protocol stacks. The Switching Module also makes intelligent interface selection in the presence of overlapped coverage between UMTS and WLAN networks. During the course of the simulation, the DMT powers up in the UMTS cell and start moving at a speed of 30 miles/hr. The DMT follows a predefined trajectory . The mobility of the DMT causes it to make handoffs from the UMTS to the WLAN Network (upon entering the WLAN network) and vice versa (upon exiting from the WLAN coverage area). The UMTS and WLAN networks are integrated at the GGSN node. The WLAN network is considered as a separate packet data network.
 

However, the WLAN coverage area is completely within the UMTS coverage area. When the DMT powers on, it first completes GPRS Attach signalling procedure signalling. The Next, DMT activates a Packet Data Protocol (PDP) context in the GGSN. A PDP context is a logical association between a mobile node and a 3G network. The PDP context defines aspects such as QoS, security, billing etc. Upon receipt of the Activate PDP Context Request, the SGSN sends a Radio Access Bearer (RAB) Assignment Request to the RNC along with the QoS requested. The UTRAN performs admission control to determine if the request can be granted. If the request can be granted, the RNC sends a Radio Bearer Setup request to the DMT.#p#分頁標題#e#
 

On receipt of the Radio Bearer Setup request, the DMT sets up the channel as specified in the request and send a Radio Bearer Complete to the RNC. When the RNC receives the Radio Bearer Complete, it sends a RAB Assignment Response, which includes the granted QoS, to the SGSN/GGSN. The DMT can send packets to the destination on receipt of the Activate PDP Context Accept message from the SGSN. Before reaching their destination, these packets are first tunnelled through the serving RNC and GGSN (using the GTP protocol), and then routed through the IP cloud to their destination node. The DMT then moves while transmitting and receiving packets using its UMTS in INTEGRATION interface. These packets are routed through the UMTS core network. At this time, the DMT's WLAN interface is in the passive scan mode and monitors the physical layer for WLAN beacon signals. Upon entering the WLAN coverage area, the WLAN interface receives beacon signals from the WLAN access point. This requires the DMT to initiate a handoff in order to switch to the WLAN access interface. The criterion for interface selection used is the bandwidth of the available access network. We have designed and implemented a Switching Module, which is really a software layer responsible for making the access interface switching decisions, and intelligently selecting an access network in the presence of overlapped coverage. The design of this Switching Module and the handoff signalling procedures are explained in the following sections. The GGSN maintains a table mapping of each subscriber's International Mobile Subscriber Identifier (IMSI) and IP Address. As the user moves across different networks, it keeps sending update messages to the GGSN about the change in its IP Address.
 

6.1 Handoff Process between UMTS and WLAN Networks

When the DMT enters the WLAN coverage area and receives a beacon signal from the WLAN AP, the WLAN_MAC layer computes the signal strength of the received signal. If the strength of the received signal is greater than the minimum packet-reception power threshold value, then the WLAN_MAC layer sends a message (which includes the BSSID [13]) to the Switching Module, indicating that a WLAN AP has been located. This implies that when the signal from the WLAN access point is not stable enough (or has a value lower than the minimum packet-reception power threshold), the handoff to the WLAN is not initiated. Thus, the case of having a weak radio signal at the boundary is accounted for. However, for a transition from the WLAN to a UMTS network, the detection of a weak WLAN signal leads to the handoff to the UMTS network since the UMTS network is considered as an always-on overlay network having a wider coverage area encompassing the WLAN.
 

If the BSSIDs match, the Switching Module sends a message to the WLAN_MAC layer to begin association with the access-point. Subsequently, the DMT attaches to the WLAN AP. The DMT continues to use both of its radio interfaces for some time: the UMTS interface continues data transmission and reception and the WLAN interface continues to send and receive signalling messages to associate with an access point. This technique is called “soft handoff” and avoids any disruption of service while switching from the UMTS to the WLAN interface. Once the DMT is completely associated with the WLAN AP, the Switching Module sends an Update Message to the GGSN node with its new IP address. Thereafter, the incoming packets are routed by the GGSN to the WLAN access point and delivered to the DMT through the WLAN interface. When the DMT moves out of coverage of the WLAN network, the WLAN interface detects the reduced strength of the received signal and passes this information to the Switching Module. The Switching Module then sends a message to the GGSN with its UMTS IP address. Thereafter, the packets are tunnelled and routed by the GGSN to the RNC, which routes them to the DMT's UMTS interface. A short handoff delay is observed when the handoff is from the WLAN to the UMTS network. This process called “hard handover” occurs when the old connection needs to be released before establishing a new connection.#p#分頁標題#e#
 

The reason for using soft handoff in the UMTS-to-WLAN switching and hard handoff in the WLAN-to-UMTS switching is because the network topology is created in such a way that the WLAN network is completely within the UMTS network. So, any mobile device that is within the WLAN network is technically also within the UMTS network.
 

Therefore, soft handoff can be used while moving from UMTS to WLAN network. In this case, the mobile device can continue to use the UMTS interface until a connection is completely established through the WLAN interface. However, while moving from WLAN network into the UMTS network, the mobile device loses its connection from the WLAN network as it moves out of the WLAN coverage area. There is a short time delay between the loss of the WLAN connection and the establishment of the UMTS connection.
 

In this case, the mobile device cannot use the WLAN connection until a UMTS connection is established since the mobile device moves out of the WLAN area before the UMTS connection is fully established. Thus it is not possible to use soft handoff in this case and hard handoff needs to be implemented.
 

6.2 Performance Evaluation

We evaluated the performance of our dual-mode radio access protocol design and the Switching Module through experimental simulations using various types of applications including: Voice over IP (VoIP) in GSM encoded format, FTP, and HTTP (web browsing). The parameters measured include end-to-end packet delay, file upload time, HTTP page response time, and handoff delay when moving from one network to another. FTP flows are cond with a constant file size of 85000 Kilobytes. Real-time VoIP flows were cond with a constant packet size of 33 bytes. Page size used for HTTP transfers was 3000 bytes.
 

Table 4 shows the response times obtained in the UMTS and the WLAN networks. In the UMTS coverage area, the FTP upload time is on an average 20.071 seconds and as the user enters into the WLAN coverage and switches transmission through the WLAN interface, the upload response time drops to an average of 0.6 seconds. We obtained similar reductions in the case of end-to-end delays obtained for voice (average of 0.348sec in UMTS network and 0.005 sec in WLAN network) and HTTP page retrieval time (average of 1.268 seconds in the UMTS network and 0.032 seconds in the WLAN network). The drop in the response times is most likely due to the higher bandwidth offered by the WLAN network eleven mega bit per second compared to one hundred and forty four kilo bit per second offered by the UMTS network.
 

The handoff delay experienced when moving across the UMTS and WLAN networks. The handoff delay values show that there is very little delay when switching the access interface while moving from UMTS to WLAN. This is because of the implementation of soft handover in which data transmission continues to take place through the UMTS interface until the mobile node is completely associated with the WLAN AP and has received a registration confirmation from the GGSN node. Then the Switching Module directs traffic through the WLAN interface. Since the WLAN network is completely within a UMTS cell, it is possible to achieve a smooth handoff when a user moves into a WLAN coverage area. However, when a user moves out of the WLAN coverage area, the WLAN interface detects a weak beacon and informs the Switching Module.#p#分頁標題#e#


This causes the Switching Module to send an IP Address update message to the GGSN. Since it takes some time for the GGSN to receive this message and update the current IP Address of the DMT, a small handoff delay is experienced. However, traffic on the UMTS uplink channel still experiences no delay since the UMTS interface is active and can be immediately used to transmit data. Packets received from the Internet are slightly delayed. However, this delay is extremely small and does not introduce a large service disruption period. Our proposed and implemented design architecture that enables seamless internetworking between UMTS and WLAN networks has several benefits. First, the WLAN can be deployed as an independent network and combined ownership of the networks is not required. Second, the approach does not require any modifications to existing 802.11 access points. Third, since all traffic passes through the central GGSN node, it is also possible to use common billing and authentication mechanisms. The low handoff involved when moving across these networks makes roaming completely transparent and seamless to the mobile users and is thus well suited to support real-time applications.
 

7. COMPARISON OF UMTS-WLAN INTERNETWORKING APPROACHES

Various approaches for integrating WLANs and 3G networks have been proposed in the literature. Some of these approaches include Mobile-IP [5], Gateway [10], Emulator [5], tight coupling, loose coupling [9], etc. However, each of these approaches integrates the WLAN and 3G networks using the integration point as either SGSN or GGSN.
 

For example, Mobile-IP and Gateway approaches are based on the GGSN-integration (using the Gi interface). The emulator approach is based on the SGSN integration (using the Gb interface). Furthermore, tight coupling is a SGSN-based approach where as Loose coupling is a GGSN-based integration scheme. Other characteristics of these approaches have been outlined in Table 1. Consequently, integrating 3G networks with WLANs requires a design that either uses SGSN or the GGSN as the integration point. Previous research efforts (Tsao et al. [5] etc.) have analyzed the performance of UMTS-WLAN integration approaches with respect to the handoff latency experienced by users while moving from one network to another. In contrast to such previous efforts, the main contributions of this work are as follows:
 

We have analyzed the two basic integration architectures (SGSN-based versus GGSN based) from a design perspective. We evaluated the performance of these integration approaches with respect to the end-to-end packet latency experienced when user traffic from WLAN and UMTS networks is routed through the SGSN (Gi interface) and the GGSN (Gb interface). The latency experienced by the packets in each case is influenced by the operations performed on the different types of WLAN access points connecting the SGSN and the GGSNs. Our results show that the end-to-end packet latency is lower when GGSN is used as the integration point as compared to the SGSN-integration due to the extra processing overheads involved at the specialized WLAN access point used in the latter case.#p#分頁標題#e#
 

We have proposed a handoff technique that achieves a lower handoff delay as compared to existing GGSN-based handoff approaches (Mobile IP, Gateway, etc.) while switching network access from 3G (UMTS) to WLANs. Our approach achieves a seamless network transition between cellular and wireless local area networks by exploiting soft handoff techniques. Fig. 6 shows a comparison of the handoff delays obtained using different types of 3G-WLAN internetworking approaches. The proposed handoff approach yielded the lowest handoff delay thereby demonstrating better performance compared to previously proposed handoff techniques.
 

We have discussed the design and implementation of a software layer called the Switching Module that enables dual-mode user terminals to perform seamless handovers across cellular and Wireless Local Area Networks using the proposed handoff technique.
 

8. CONCLUSION

Integrated WLAN and UMTS networks benefit users with both high-speed connectivity as well as widespread coverage. Development of architectures that allow interoperability and internetworking between these technologies along with seamless roaming facility is a challenge today. In this work, we have proposed, implemented and evaluated architecture for a dual interface mobile node that implements a dynamic interface switching algorithm based upon the received signal strength of the WLAN beacon signal.
 

The technique enables smooth handoff process to take place while moving across heterogeneous networks such as UMTS and WLANs. The performance evaluation of the architecture shows minimal handoff delay while switching from UMTS to WLAN networks
 

And a slightly higher handoff delay incurred when moving out of the WLAN coverage and switching radio access to the UMTS interface. The small delays obtained make this architecture feasible for use in both real-time as well as non real-time environments. Another feature of our design approach is that it is based on simple IP services and does not require the use of Mobile IP or IPv6 mechanisms for seamless connectivity.
 

REFERENCES

1. IEEE 802.11b, Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications-Higher-Speed Physical Layer Extension in the 2.4 GHz Band, 1999.

2. H. Luo, Z. Jiang, B. J. Kim, N. K. Shankar, and P. Henry, “Internet roaming: a WLAN/3G integration system for enterprises,” AT&T Labs-Research, http://www.nd.edu/~hluo/publications/SPIE02.pdf.

3. J. D. Vriendt, P. Laine, C. Lerouge, and X. Xiaofeng, “Mobile network evolution: a revolution on the move,” IEEE Communications Magazine, Vol. 40, 2002, pp. 104-111.

4. TS 125 422-V6.0.0, “Universal mobile telecommunications system,” http://www.hwebapp.etsi.org/action%5CPU/20040120/ts_125422v060000p.pdf.

5. S. Tsao and C. Lin, “Design and evaluation of UMTS-WLAN internetworking strategies,” in Proceedings of IEEE Vehicular Technology Conference, Vol. 2, 2002, pp.777-781.#p#分頁標題#e#

6. M. Buddhikot, G. Chandranmenon, S. Han, Y. Lee, and S. Miller, “Design and implementation of a WLAN/CDMA2000 internetworking architecture,” IEEE Communications Magazine, 2003, pp. 90-100.

7. B. Hongyang, H. Chen, and L. Jiang, “Performance analysis of vertical handoff in a UMTS-WLAN integrated network,” in Proceedings of the 14th IEEE Conference on Personal, Indoor and Mobile Radio Communications, Vol. 1, 2003, pp. 187-191.

8. P. Matusz, P. Machan, and J. Wozniak, “Analysis of profitability of inter-system handoffs between IEEE 802.11b and UMTS,” in Proceedings of the 28th Annual IEEE International Conference on Local Computer Networks, 2003, pp. 210-217.

9. A. Salkintsiz, C. Fors, and R. Pazhyannur, “WLAN-GPRS integration for next-generation mobile data networks,” IEEE Wireless Communications, Vol. 9, 2002, pp.112-124.

10. S. Tsao and C. Lin, “VSGN: a gateway approach to interconnect UMTS/WLAN networks,” in Proceedings of the 13th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, Vol. 1, 2002, pp. 275-279.

11. OPNET Inc., http://www.opnet.com.

12. V. Varma, S. Ramesh, M. Barton, G. Hayward, and J. A. Friedhoffer, “Mobility management in UMTS/WLAN networks,” IEEE International Conference on Communications, Vol. 2, 2003, pp. 1048-1053.

13. S. Uskela, “Key concepts for evolution toward beyond 3G networks,” IEEE Wireless Communications, Vol. 10, 2003, pp. 43-48.

Appendix Data Table Of Simulation

Table 4 (1UE vs 1UW)

UMTS UE FTP Download Response Time with FTP File size

File size

1B

50KB

100KB

200KB

500KB

700KB

900KB

1MB

1.5MB

1.8MB

2MB

3MB

Seed 1

1.252

21.943

37.783

66.343

182.963

238.849

390.563

370.063

533.349

653.749

724.449

1053.533

Seed 2

1.252

37.369

38.849

61.249

179.069

250.249

344.869

360.849

538.449

646.435

731.235

1071.933

Seed 3

1.253

47.949

38.649

73.249

179.849

252.533

323.049

364.249

561.335

630.633

716.333

1063.932

Seed 4

1.253

27.249

37.749

70.855

176.855

259.273

326.749

363.055

556.533

665.433

711.533

1082.632

Seed 5

1.253

27.835

39.735

77.273

178.653

249.252

324.735

368.733#p#分頁標題#e#

554.533

651.749

743.053

1061.932

Mean

1.253

32.469

38.553

69.794

179.478

250.031

341.993

365.390

548.840

649.600

725.321

1066.792

95% CI:

1.252

19.700

37.527

62.093

176.701

240.880

306.571

360.566

533.690

633.868

709.851

1053.123

1.253

45.238

39.579

77.495

182.255

259.183

377.415

370.214

563.989

665.331

740.790

1080.462

UW (WLAN Mode)

File size

1B

100KB

250KB

800KB

1.5MB

4MB

7MB

12MB

15MB

20MB

30MB

40MB

Seed 1

0.195

1.178

2.358

6.338

11.718

30.278

52.458

89.958

112.518

149.798

223.938

298.058

Seed 2

0.189

1.191

2.171

6.351

11.731

30.291

52.471

89.971

112.531

149.831

223.931

298.071

Seed 3

0.184

1.204

2.184

6.364

11.744

30.369

52.484

89.784

112.409

149.624

223.744

297.884

Seed 4

0.191

1.069

2.249

6.429

11.609

30.304

52.549

89.849

112.344

149.715

223.855

297.955

Seed 5

0.198

1.095

2.275

6.455

11.635

30.395

52.575

89.875

112.435

149.689

223.809

297.949

Mean:

0.191

1.147

2.247

6.387

11.687

30.327

52.507

89.887

112.447

149.731

223.855

297.983

95% CI:

0.184

1.071

2.153

6.323

11.611

30.263

52.443

89.790

112.350

149.627

223.753

297.885

0.198

1.223

2.341

6.451

11.763

30.391

52.571

89.984

112.544

149.835

223.957

298.081

Table 5

UMTS UEs Number of Clients Vs FTP Download Response Time

Number of UEs

30

28

26

24

22

20#p#分頁標題#e#

18

16

Seed 1

39.970

30.263

24.622

22.263

20.563

19.963

19.414

19.363

Seed 2

39.852

30.539

24.516

22.305

20.586

19.931

19.361

19.480

Seed 3

39.955

30.540

24.534

22.358

20.575

19.915

19.399

19.438

Seed 4

39.954

30.529

24.520

22.173

20.586

19.947

19.450

19.403

Seed 5

39.852

30.527

24.527

22.292

20.527

19.937

19.527

19.377

Mean:

39.916

30.480

24.544

22.278

20.567

19.939

19.430

19.412

95% CI:

39.843

30.329

24.489

22.194

20.537

19.917

19.352

19.353

39.990

30.630

24.598

22.363

20.597

19.961

19.508

19.471

No of users

(COUNT)

14

12

10

08

06

04

02

Seed 1

19.263

19.163

19.063

18.663

18.283

18.279

18.275

Seed 2

19.323

19.314

19.004

18.560

18.288

18.330

18.322

Seed 3

19.235

19.262

19.004

18.645

18.317

18.255

18.314

Seed 4

19.329

19.162

19.130

18.934

18.280

18.270

18.288

Seed 5

19.231

19.160

19.022

18.726

18.338

18.274

18.274

Mean:

19.276

19.212

19.045

18.705

18.301

18.282

18.294

95% CI:

19.218

19.123

18.978

18.531

18.270

18.247

18.267

19.335

19.301

19.111

18.880

18.332

18.317

18.322

(WLAN Mode) WLAN_UMTS UWs

No of UWs

30

28

26

24

22

20

18

16

14

Seed 1

0.5326

0.5327

0.5358

0.5418

0.5383

0.5326

0.5268

0.5217

0.5358

Seed 2

0.5323

0.5268

0.5420

0.5508

0.5470

0.5323

#p#分頁標題#e#

0.5327

0.5225

0.5420

Seed 3

0.5200

0.5348

0.5384

0.5447

0.5344

0.5200

0.5268

0.5266

0.5384

Seed 4

0.5271

0.5348

0.5568

0.5409

0.5198

0.5271

0.5348

0.5267

0.5568

Seed 5

0.5235

0.5229

0.5253

0.5394

0.5454

0.5235

0.5229

0.5251

0.5253

Mean:

00.527

00.530

00.539

00.543

00.537

00.527

00.528

00.524

00.539

95%CI:

0.2503

0.2538

0.5255

0.5379

0.5234

0.5203

0.5228

0.5216

0.5255

0.5339

0.5370

0.5539

0.5491

0.5505

0.5339

0.5348

0.5275

0.5539

No of UWs

(COUNT)

12

10

08

06

04

02

Seed 1

0.5418

0.5383

0.5430

0.5629

0.5428

0.5960

Seed 2

0.5508

0.5470

0.5402

0.5469

0.5858

0.5567

Seed 3

0.5447

0.5344

0.5460

0.5617

0.5273

0.5224

Seed 4

0.5409

0.5198

0.5348

0.5465

0.5353

0.5395

Seed 5

0.5394

0.5454

0.5558

0.5684

0.5481

0.5224

Mean:

0.5435

0.5370

0.5440

0.5573

0.5479

0.5474

95%CI:

0.5379

0.5234

0.5343

0.5449

0.5198

0.5093

0.5491

0.5505

0.5536

0.5697

0.5759

0.5855

Table 6 (UEs vs UWs)

(UEs vs UWs) No of Clients Vs Web Page Response Time

No of UEs

30

28

26

24

22

20

18

Seed 1

19.426

16.060

13.760

12.456

11.559

10.599

9.959

Seed 2

19.438

16.140

13.641

12.538

11.739

10.354

10.034

Seed 3

19.443

16.041

13.837

12.742

11.651

10.436

9.954

Seed 4

19.535

16.239

13.738

12.641

11.539

10.541

9.862

Seed 5

19.739

16.141

13.643#p#分頁標題#e#

12.838

11.541

10.545

10.039

Mean:

19.516

16.124

13.724

12.643

11.606

10.495

9.969

95%CI:

19.353

16.027

13.621

12.453

11.497

10.373

9.880

19.680

6.222

13.827

12.833

11.715

10.617

10.059

No of UEs

16

14

12

10

08

06

04

02

Seed 1

9.585

9.185

8.685

8.285

8.184

8.144

8.084

8.004

Seed 2

9.638

9.253

8.660

8.347

8.195

8.174

8.086

8.098

Seed 3

9.543

9.244

8.748

8.353

8.177

8.157

8.089

8.049

Seed 4

9.551

9.153

8.753

8.344

8.161

8.139

8.092

8.056

Seed 5

9.641

9.141

8.651

8.338

8.209

8.154

8.080

8.031

Mean:

9.592

9.195

8.699

8.334

8.185

8.154

8.086

8.048

95%CI:

9.534

9.132

8.639

8.299

8.163

8.137

8.080

8.005

9.649

9.259

8.759

8.368

8.208

8.171

8.092

8.091

No of UWs

30

28

26

24

22

20

18

16

Seed 1

0.2409

0.2258

0.2352

0.2384

0.2444

0.2490

0.2285

0.2352

Seed 2

0.2420

0.2401

0.2389

0.2418

0.2456

0.2420

0.2401

0.2389

Seed 3

0.2367

0.2499

0.2360

0.2409

0.2453

0.2367

0.2299

0.2360

Seed 4

0.2453

0.2412

0.2411

0.2384

0.2444

0.2453

0.2412

0.2411

Seed 5

0.2328

0.2339

0.2397

0.2428

0.2425

0.2382

0.2393

0.2397

Mean:

00.240

00.237

00.238

00.240

00.244

00.240

00.237

00.238

95%CI:

0.2317

0.2239

0.2351

0.2380

0.2430

0.2371

0.2293#p#分頁標題#e#

0.2351

0.2430

0.2455

0.2413

0.2429

0.2459

0.2430

0.2455

0.2413

No of 14 12 10 08 06 04 02

UWs

Seed 1 0.2384 0.2444 0.2395 0.2450 0.2493 0.2479 0.2510

Seed 2 0.2418 0.2456 0.2409 0.2451 0.244 0.2496 0.2513

Seed 3 0.2409 0.2453 0.2410 0.2440 0.2443 0.2428 0.2432

Seed 4 0.2384 0.2444 0.2483 0.2392 0.2428 0.2476 0.2619

Seed 5 0.2428 0.2425 0.2470 0.2511 0.2527 0.2494 0.2432

95% CI: 0.2380 0.2430 0.2383 0.2396 0.2417 0.2440 0.2405

Mean: 00.240 00.244 00.243 00.244 00.246 00.247 00.250

00.242 00.245 00.248 00.250 00.251 00.250 00.259

Table 7 (WLAN vs UW)

WLAN Client FTP Download Response Time vs FTP File size

File size 1B 100KB 250KB 800KB 1.5MB 4MB 7MB 12MB

Seed 1 0.195 1.178 2.358 6.338 11.718 30.278 52.458 89.958

Seed 2 0.189 1.191 2.171 6.351 11.731 30.291 52.471 89.971

Seed 3 0.184 1.204 2.184 6.364 11.744 30.369 52.484 89.784

Seed 4 0.191 1.069 2.249 6.429 11.609 30.304 52.549 89.849

Seed 5 0.198 1.095 2.275 6.455 11.635 30.395 52.575 89.875

Mean: 0.191 1.147 2.247 6.387 11.687 30.327 52.507 89.887

95% CI: 0.184 1.071 2.153 6.323 11.611 30.263 52.443 89.790

0.198 1.223 2.341 6.451 11.763 30.391 52.571 89.984

FTP File size 15MB 20MB 30MB 40MB 60MB 80MB 100MB

Seed 1 112.518 149.798 223.938 298.058 448.738 599.358 749.735

Seed 2 112.531 149.831 223.931 298.071 448.751 599.371 749.709

Seed 3 112.409 149.624 223.744 297.884 448.564 599.184 749.644

Seed 4 112.344 149.715 223.855 297.955 448.609 599.249 749.791

Seed 5 112.435 149.689 223.809 297.949 448.635 599.275 749.778

Mean: 112.447 149.731 223.855 297.983 448.659 599.287 749.731

95% CI: 112.350 149.627 223.753 297.885 448.557 599.190 749.658

112.544 149.835 223.957 298.081 448.761 599.384 749.804

UW (WLAN Mode)

File size 1B 100KB 250KB 800KB 1.5MB 4MB 7MB 12MB

Seed 1 0.042 0.395 0.770 2.010 3.410 8.850 15.272 25.990

Seed 2 0.026 0.363 0.737 1.977 3.377 8.817 15.240 25.957

Seed 3 0.027 0.331 0.705 1.945 3.345 8.824 15.208 25.925

Seed 4 0.040 0.369 0.744 1.984 3.384 8.785 15.246 25.964

Seed 5 0.032 0.505 0.679 1.919 3.519 8.759 15.182 25.899

Mean: .033 0.393 0.727 1.967 3.407 8.807 15.229 25.947

95% CI: 0.025 0.310 0.683 1.923 3.324 8.763 15.186 25.903

0.042 0.475 0.771 2.011 3.490 8.851 15.273 25.991

File size 15MB 20MB 30MB 40MB 60MB 80MB 100MB

Seed 1 32.410 43.112 64.530 85.950 128.790 171.604 214.359

Seed 2 32.377 43.080 64.497 85.917 128.757 171.572 214.424

Seed 3 32.384 43.048 64.465 85.885 128.725 171.540 214.385

Seed 4 32.345 43.022 64.439 85.859 128.764 171.579 214.417

Seed 5 32.319 43.086 64.504 85.924 128.699 171.514 214.450#p#分頁標題#e#

Mean: 32.367 43.070 64.487 85.907 128.747 171.562 214.407

95% CI: 32.323 43.026 64.443 85.863 128.703 171.518 214.363

32.411 43.113 64.531 85.951 128.791 171.606 214.451

Table 8

WLAN Mode (UWs)

No of UW 0 2 4 6 8 10 12 14 16 18 20

Seed 1 0.251 0.248 0.249 0.245 0.239 0.244 0.242 0.241 0.241 0.239

Seed 2 0.251 0.250 0.245 0.245 0.241 0.246 0.241 0.240 0.240 0.241

Seed 3 0.243 0.243 0.244 0.244 0.241 0.245 0.243 0.241 0.241 0.240

Seed 4 0.262 0.248 0.243 0.251 0.248 0.244 0.241 0.240 0.239 0.244

Seed 5 0.243 0.249 0.253 0.251 0.247 0.243 0.243 0.241 0.241 0.241

Mean: 00.25 00.24 00.24 00.24 00.24 00.24 00.24 00.24 00.24 00.24

95% CI: 0.241 0.244 0.242 0.243 0.238 0.243 0.241 0.240 0.239 0.239

0.260 0.251 0.252 0.252 0.248 0.246 0.243 0.241 0.242 0.243

Table 9

UMTS Mode (UWs) No of Clients Vs FTP Download Response Time

No of UWs 0 2 4 6 8 10 12 14 16 18 20

Seed 1 0.596 0.543 0.563 0.543 0.538 0.542 0.536 0.529 0.533 0.533

Seed 2 0.557 0.586 0.547 0.540 0.547 0.551 0.542 0.525 0.533 0.533

Seed 3 0.540 0.527 0.562 0.546 0.534 0.545 0.538 0.529 0.527 0.532

Seed 4 0.557 0.535 0.547 0.535 0.545 0.541 0.557 0.529 0.535 0.527

Seed 5 0.540 0.548 0.568 0.556 0.545 0.539 0.525 0.525 0.523 0.524

Mean: 00.55 00.54 0 0.55 0 0.54 00.54 00.54 00.54 00.52 00.53 00.53

95% CI: 0.529 0.520 0.545 0.534 0.535 0.538 0.525 0.525 0.524 0.524

0.586 0.576 0.570 0.554 0.549 0.549 0.554 0.530 0.536 0.535

TABLE 10

(Heavy Traffic Load) Number of UWs Vs Web Page Response Time

No of UWs 2 6 10 16 20 23 25 28 30

Seed 1 1.140 1.347 1.516 1.538 1.950 2.718 4.546 7.947 15.663

Seed 2 1.298 1.357 1.347 1.343 1.754 2.649 4.546 8.355 15.565

Seed 3 1.349 1.439 1.353 1.651 1.836 2.789 4.160 8.177 15.734

Seed 4 1.356 1.354 1.444 1.441 1.941 2.860 4.548 7.961 15.854

Seed 5 1.256 1.239 1.738 1.832 1.945 2.757 4.526 7.965 15.562

Mean: 1.280 1.353 1.480 1.561 1.885 2.754 4.466 8.081 15.676

95% CI: 1.170 1.263 1.281 1.325 1.776 2.657 4.254 7.857 15.523

1.390 1.442 1.679 1.797 1.994 2.852 4.678 8.305 15.828