6
Network architecture supporting
wireless applications
The Wireless Application Environment (WAE) architecture is designed to support Mobile
Terminals (MTs) and network applications using different languages and character sets.
WAE user agents have a current language and a ccept content in a set of well-known
character encoding sets. Origin server-side applications can emit content in one or more
encoding sets and can accept input from the user agent in one or more encoding sets.
Wireless Telephony Application (WTA) is an application framework for telephony ser-
vices. The WTA user agent has the capability for interfacing with mobile network services
available to a mobile telephony device, that is, setting up and receiving phone calls.
The Wireless Application Protocol (WAP) Push framework introduces a means within
the WAP effort to transmit information to a device without a previous user action. In
the client/server model, a client requests a service or information from a server, which
transmits information to the client. In this pull technology, the client pulls information
from the server.
6.1 WAE ARCHITECTURE
The WAE architecture includes networking schemes, content formats, programming lan-
guages, and shared services. Interfaces are not standardized and are specific to a particular
implementation. WAE can work with a browser and a class of user agents used in the
World Wide Web (WWW).
In the Internet WWW, applications present content to a client in a set of standard
data formats that are browsed by client side user agents known as Web browsers.Auser
agent sends requests for one or more data objects or content to an origin server, which
responds with the requested data expressed in one of the standard formats known to
the user agent [i.e., Hypertext Markup Language (HTML)]. The WWW logical model is
shown in Figure 6.1.
Mobile Telecommunications Protocols For Data Networks. Anna Ha
´
c
Copyright

Response (content)
Figure 6.2 WAE logical model.
All resources on the WWW are named using Internet standard Uniform Resource Loca-
tors (URLs). All classes of data on the WWW are given as specific types, allowing the
user agent to correctly distinguish and present them appropriately. The WWW defines
a variety of standard content formats supported by most browser user agents, including
the HTML, the JavaScript scripting language, and other formats like bitmap image for-
mats. The WWW defines a set of standard networking protocols allowing any browser to
communicate with any origin server, for example, Hypertext Transfer Protocol (HTTP).
The WAE logical model is shown in Figure 6.2. In the WAE model, the content is
transported using standard protocols in the WWW domain a nd an optimized HTTP-
like protocol in the wireless domain. The content a nd services in WAE architecture are
hosted on standard Web origin servers using proven technologies like Common Gateway
Interface (CGI). The content is located by using WWW standard URLs. WAE supports
WAE ARCHITECTURE 95
Mobile Network Services such as Call Control and Messaging. WAE architecture supports
low bandwidth and high latency networks and considers CPU processing constraints in
MTs. WAE assumes the existence of gateway functionality responsible for encoding and
decoding data transferred from and to the mobile client. The purpose of the encoding
content delivered to the client is to minimize the size of data sent to the client Over
The Air (OTA), and to minimize the computational energy required by the client to
process the data. The gateway functionality can be added to origin servers or placed in
dedicated gateways.
The main elements of the WAE model are WAE user agents, content generators,
standard content encoding, and WTA. WAE user agents interpret network content ref-
erenced by a URL. Content generators are the applications or services on origin servers,
like CGI scripts, that produce standard content formats in response to requests from user
agents in MTs. Standard content encoding allows a WAE user agent to navigate Web
content. WTA is a collection of telephony-specific extensions for call and feature control
mechanisms providing advanced Mobile Network Services.

96 NETWORK ARCHITECTURE SUPPORTING WIRELESS APPLICATIONS
WAE
User agents
Services / formats
WTA user agent
WML user agent
Other agents
Other
applications
and
services
WMLScript
WAP Protocol stack and services
Device OS/services
WTA srvs
URLs
WML
Other srvs.
& formats
Figure 6.3 WAE client components.
limited user input facilities, narrow-band network connections, limited memory resources,
and limited computational resources.
The WML features include
• support for text and images;
• support for user input;
• navigation and history stack;
• international support;
• Man–Machine Interface (MMI) independence;
• narrow-band optimization; and
• state and context management.

Figure 6.4 WML user agent logical architecture with gateway.
scheme operation. The origin server replies by sending a single deck in a textual format.
On their way back to the client, textual decks are expected to pass through a gateway
where they are converted into formats better suited for OTA transmission and limited-
device processing. The gateway does all the necessary conversions between the textual
and binary formats. A WML encoder (or tokenizer) in the gateway converts each WML
deck into its binary format. Encoded content is then sent to the client to be displayed and
interpreted.
The user agent may submit one or more additional requests, using a URL scheme, for
WMLScript as the user agent encounters references to them in a WML deck. On its way
back, a WMLScript compiler takes the script as input and compiles it into byte code that
is designed for low bandwidth and thin mobile clients. The compiled byte code is then
sent to the client for interpretation and execution.
Figure 6.5 shows W ML user agent logical architecture without a gateway. WAE does
not specify the location where the actual e ncoding and compilation is done. The origin
servers may have built-in WML encoders and WMLScript compilers. Some services may
Client
Origin server
CGI
scripts,
etc.
Content
WML
user
agent
WAE
services
WML encoder
WMLScript
compiler

and signaling
Mobile
network
WTA
server
Other
applications
Direct user
interactions
WTAI libraries
Mobile
client
WAP
gateway
Figure 6.6 WTA architecture.
WTA ARCHITECTURE 99
WAE
user agent
Mobile
client
WAP
gateway
Firewall
(optional)
Internet
Mobile
network
Man−machine interface
Device-specific features Network layer
Network events

2
3
2
3
1a
1a
1b
1b
Access to a URL (via the repository)
Access to a URL (via the WTA server)
Service Indication (Push)
Network event (transformed to WTA event in client)
WAP
gateway
Mobile
network
WTA
server
Figure 6.8 Initiation of WTA services.
on a WTA server provides a simple and powerful model on how to seamlessly integrate
services in the mobile network with services executing locally in the WAP client.
WTA services appear to the client in the form of various content formats, such as
WTA–WML, WMLScript, and so on. The W TA user agent executes content that is
persistently stored in the client’s repository or content retrieved from a WTA server. The
WTA user agent can act on events from the mobile network, for instance, an incoming call.
Figure 6.8 shows how to initiate a WTA service in the WTA user agent. The WTA user
agent executes content within the boundary of a well-known context. The service defines
the extent of a context and its associated content. The start of a service is marked by the
initiation of a new context, and the termination of a context marks the end of a service.
The repository is a persistent storage module within the MT that may be used to

mobile client. These functions allowing for setting up calls, and accessing the users local
phone book, must ensure that only authorized WTA services are permitted to execute. The
trusted mobile telephony service provider, which provides an acceptable level of security
in the network, can choose to run all WTA services itself not allowing other providers
or it can choose to delegate the administration of its WTA services to a third party. The
Wireless Datagram Protocol (WDP) uses predefined port numbers to separate a WTA
service from a common WAE service as shown in Figure 6.9. A WTA session established
by the WTA user agent must use one of the dedicated, secure WTA ports on the gateway.
The WTA user agent must not retrieve WTA content outside the WTA session. WTA
content received outside the WTA session and Service Indication addressing the WTA
user agent but delivered outside a WTA session shall be discarded.
The repository is used to store WTA content persistently. This provides a mechanism
that ensures timely handling of content related to WTA services initiated by WTA events
and has the following characteristics:
• The repository contains a set of channels and resources.
• Resources are data downloaded with WSP (that is, WTA–WML deck) and are stored
along with their metadata, that is, content type and the HTTP 1.1 entity tag, and
location (URL).
• A channel is a resource that contains a set of links and resources and has identity
and freshness.
• Channels in the repository have a freshness lifetime (the HTTP 1.1 expiry date header),
beyond which time they are considered stale. Stale channels are subject to automatic
removal by the user agent. Resources are subject to automatic removal from the repos-
itory if the channel does not reference them.
• If the r epository contains a channel that is not stale, it is guaranteed that the repository
contains all resources named in that channel. The loading and unloading of a channel
is an atomic operation in that no user agent will recognize the presence of the channel
until all the content in the channel has been successfully stored in the repository.
102 NETWORK ARCHITECTURE SUPPORTING WIRELESS APPLICATIONS
Repository

them from the server. The list is presented on the client’s display. When a certain voice
mail has been selected, the server sets up a call to the client and the user listens to the
selected voice mail.
WTA ARCHITECTURE 103
• Call subscriber from message list or log: When a list of voice, fax, or e-mails or any
kind of call log is displayed, the user has the option of calling the originator of a
selected entry in the list or log.
The incoming call selection service is started when an incoming call is detected in
the client and a menu with various call-handling options is presented to the user as
shown in Figure 6.11. A valid channel and its associated content are stored in the repos-
itory. The client is not engaged in any other WTA service (i.e., no temporary event
bindings exist).
The following events in this example of incoming call selection are shown in Figure 6.11.
1. The mobile network r eceives an incoming call and sends a Call Indication to the
mobile subscriber.
2. In the client, the incoming call WTA event (wtaev-cc/ic) is generated. The repository
is checked to find a dedicated channel. The channel provides the URL to the Incoming
Call Selection service stored in the repository.
3. The user agent requests the content from the repository.
4. The repository returns the requested content.
5. The content is loaded into a clean context and starts executing. The service presents to
the user a list of options, from which he or she can choose how to proceed with a call
in progress. In this example, the user elects to answer the call. The WTAI function
WTAIVoiceCall.accept is invoked.
Answer
Hold
Reject
Voice Mail
Forward to office
>

municates with the mobile network).
7. A Connect Acknowledgement (ACK) is generated in the mobile network. A result
code indicating the outcome of the calls is generated internally in the phone.
8. A speech path between the mobile network and the client is established.
Figure 6.12 shows how the voice mail service is established w ithin the WTA frame-
work. The user is notified that he or she has received new voice mails, and the user
chooses to listen to one of them.
The following events in this example of voice mail are shown in Figure 6.12.
1. The Voice Mail System notifies the W TA server that there are new voice mails. A
list of voice mails is also sent to the WTA server.
2. The WTA server creates new service content on the basis of the list received from
the voice mail system. The content is stored on the server and its URL is included
in an SI that will be pushed to the client. The SI’s message reads: ‘You have 3 new
voice mails’.
3. The WAP gateway sends the SI to the client using push.
5
3 2
1
7
12
15
17
16
19
20
6
8
11
13
4

Push
Content
download
Content
download
WAP
gateway
WTA
server
Mobile
network
Voice
mail
New voice mail
Retrieve message (GET URL)
Call indication
Play message
Service Indication
Hello
Figure 6.12 Voice mail.
WAP PUSH ARCHITECTURE 105
4. The user is notified about the SI by a message delivered with the SI. The user chooses
to accept the SI.
5. A WSP Get request is sent to the WAP gateway (URL provided by the SI).
6. The WAP gateway makes a WSP/HTTP conversion.
7. The WTA server returns the earlier created voice mail service.
8. The WAP gateway makes an HTTP/WSP conversion.
9. The voice mail service is now executing in the client. The user is presented with a
list of voice mails originating from the Voice Mail System (a WTA–WML Select
List is created in Step 2). The user selects a certain voice mail to listen to.

technology is WWW, in which a user enters a URL (the request), sent then to a server,
which answers by sending a Web page (the response) to the user. In the push technology
based on client/server model, there is no explic it request from the client before the server
106 NETWORK ARCHITECTURE SUPPORTING WIRELESS APPLICATIONS
‘Pull’ technology
‘Push’ technology
ServerClient
Figure 6.13 Comparison of pull and push technology.
transmits its content. Figure 6.13 illustrates pull and push technology. Pull transactions
are initiated by the client, whereas push transactions are initiated by the server.
A push operation in WAP occurs when a Push Initiator transmits content to a client
using either the Push OTA protocol or the Push Access Protocol (PAP). The Push Initiator
does not share a protocol with the WAP client since the Push Initiator is on the Internet
and the WAP client is on the WAP domain. The Push Initiator contacts the WAP Client
through a translating Push Proxy Gateway (PPG) from the Internet side, delivering content
for the destination client using Internet protocols. The PPG forwards the pushed content
to the WAP domain, and the content is then transmitted over the air in the mobile network
to the destination client. The PPG may be capable of notifying the Push Initiator about
the final outcome of the push operation, and it may wait for the client to accept or reject
the content in two-way mobile networks. It may also provide the Push Initiator with
client capability lookup services by letting a Push Initiator select the optimal content for
this client.
The Internet side PPG access protocol is called the PAP. The WAP side protocol is
called OTAProtocol. The PAP uses XML messages that may be tunneled through various
Internet protocols, for example, HTTP. The OTA protocol is based on WSP services. The
Push framework with the protocols is shown in Figure 6.14.
The PPG acts as an access point for content pushes from the Internet to the mobile
network, and associated authentication, security, client control, and so on. The PPG owner
decides the policies about who is able to gain access to the WAP network, who is able to
push content, and so on. The PPG functionality may be built into the pull WAP gateway

The PPG may encode WAP content types into their binary counterparts. This transaction
takes place before delivery over the air. Other content types may be forwarded as received.
The P ush Initiator may also precompile its content into binary form to take workload off
the PPG, for example. When the PPG receives precompiled WML, WMLScript, or SIs,
they are forwarded as received.
The PPG may implement addressing aliasing schemes to enable special multi- and
broadcast cases, in which special addresses may translate to a broadcast operation.
A Push Initiator may query the PPG for client capabilities and preferences to create
better formatted content for a particular WAP device.
The PAP is used by an Internet-based Push Initiator to push content to a mobile network
addressing its PPG. The PAP initially uses HTTP, but it can be tunneled through any other
or future Internet protocol. The PAP carries an XML-style entity that may be used with
other components in a multipart-related document.
The PAP supports the following operations:
• Push Submission (Initiator to PPG)
• Result Notification (PPG to Initiator)
• Push Cancellation (Initiator to PPG)
• Status Query (Initiator to PPG)
• Client Capabilities Query (Initiator to PPG).
The push message contains three entities: a control entity, a content entity, and option-
ally a capability entity. They are used in a multipart-related message, which is sent from
the Push Initiator to the PPG. The control entity is an XML document containing delivery
instructions destined for the PPG, and the content entity is destined for the mobile device.
If the Push Initiator requested a confirmation of successful delivery, the message is
transmitted from the PPG to the Push Initiator when the content is delivered to the mobile
device over a two-way bearer, or transmitted to the device over a one-way bearer, and it
108 NETWORK ARCHITECTURE SUPPORTING WIRELESS APPLICATIONS
contains an XML entity. The message is also transmitted in case of a detected delivery
failure to inform the Initiator about it.
The Push Initiator relies on the response from the PPG; a confirmed push is then

there are no active sessions to that client, the server cannot deliver the push content.
A session request is sent to a special application in the client known as the SIA. This
request contains information necessary for a client to create a push session. The SIA
in the client after receiving a session request establishes a session with the PPG and
indicates which applications accept content over the newly opened session. The SIA may
also ignore the request if there is no suitable installed application as requested in the
session request.
When a client receives pushed content, a dispatcher looks at the push message header
to determine its destination application. This dispatcher is responsible for rejecting con-
tent that does not have a suitable destination application installed, and for confirming
push operations to the PPG when the appropriate application takes responsibility for
pushed content.
PROBLEMS TO CHAPTER 6 109
6.4 SUMMARY
The main elements of the WA E model are WAE user agents, content generators, standard
content encoding, and wireless telephony applications. WAE user agents interpret network
content referenced by a URL. Content generators are the applications or services on origin
servers, like CGI scripts, that produce standard content formats in response to requests
from user agents in MTs. Standard content encoding allows a WAE user agent to navigate
Web content. WTA is a collection of telephony-specific extensions for call and feature
control mechanisms providing advanced Mobile Network Services.
The repository is a persistent storage module within the MT that may be used to
eliminate the need for network access when loading and executing frequently used WTA
services. The repository also addresses the issue of how a WTA service developer ensures
that time critical WTA events are handled in a timely manner. The repository addresses
the issues of how the WTA services developer preprogram the device with content and
how the WTA services developer improves the response time for a WTA service.
A push operation in WAP occurs when a Push Initiator transmits content to a client
using either the Push OTA protocol or the PAP. The Push Initiator does not share a
protocol with the WAP client since the Push Initiator is on the Internet and the WAP

to requests from user agents in MTs. Standard content encoding allows a WAE
user agent to navigate Web content. WTA is a collection of telephony specific
extensions for call and feature control mechanisms providing advanced Mobile Net-
work Services.
6.2: The repository is a persistent storage module within the MT that may be used to
eliminate the need for network access when loading and executing frequently used
WTA services. The repository also addresses the issue of how a WTA service devel-
oper ensures that time-critical WTA events are handled in a timely manner. The
repository addresses the issues of how the WTA services developer preprogram the
device with content, and how the WTA services developer improves the response
time for a WTA service.
6.3: The WAP Push framework introduces a means within the WAP effort to transmit
information to a device without a previous user action. In the client/server model, a
client requests a service or information from a server, which transmits information
to the client. In this pull technology, the client pulls information from the server. An
example of pull technology is WWW, in which a user enters a URL (the request) sent
then to a server, which answers by sending a Web page (the response) to the user. In
the push technology based on the client/server model, there is no explicit request from
the client before the server transmits its content. Pull transactions are initiated by
the c lient, whereas push transactions are initiated by the server. A push operation in
WAP occurs when a Push Initiator transmits content to a client using either the Push
OTA protocol or the PAP. The Push Initiator does not share a protocol with the WAP
client since the Push Initiator is on the Internet and the WAP client is on the WAP
domain. The Push Initiator contacts the WAP Client through a translating PPG from
the Internet side, delivering content for the destination client using Internet protocols.