WAP

In 1997, the U.S. operator Omnipoint decided to roll out a mobile Web service. It had no idea how to go about this or which technology to deploy, so opened the process to competitive tender. Anyone who could come up with a wireless data proposal was free to submit it to Omnipoint, and to explain how it would serve the company's customers and increase revenue.

Four companies eventually took up this challenge. The cellphone vendors Nokia, Ericsson, and Motorola all suggested their own variants on messaging, while Unwired Planet submitted HDML. All had their advantages, but the big disadvantage of being proprietary: Omnipoint's customers would be locked into buying phones and software from a particular supplier, similar to the way that most PC users are locked into Microsoft.

Omnipoint told the bidders that it would not accept a proprietary solution—they had to get together and thrash out a standard. The result was the formation of the WAP Forum, which originally consisted of these four companies, though not Omnipoint itself. A year later, they threw the Forum open to new members and announced the first version of the Wireless Application Protocol.

WAP isn't the only nonproprietary solution. C-HTML is also open to all, and endorsed by the W3C. However, WAP differs from other wireless Web systems in that it isn't just a markup language—it's a complete new stack of protocols designed to overcome some of a wireless network's specific problems, such as high latency and jitter. This makes it more complicated than C-HTML, but also more reliable, if the WAP Forum is to be believed.

If previous battles are anything to go by, the standards war between WAP and C-HTML will depend on which has the heaviest backers, not necessarily the best technology. Here WAP has a clear advantage: it was designed in part by the mobile phone manufacturers, who still call the shots in the wireless world. A compromise may be possible: the future version 2.0 of WAP will include C-HTML, while Microsoft already has a browser than can display both.

Protocols

In networking, a protocol is a set of rules for communication between similar devices. They can cover anything, but in general protocols regulate such conditions as whose turn it is to transmit, how errors are detected or fixed, and how to distinguish data from signals sent alongside it.

Protocols were originally designed separately by dominant companies such as IBM and Novell. This meant that it was very difficult to get different vendors' devices to communicate with each other. It was as if they were not only speaking different languages, but using different vocal chords to produce them. To get around this, the International Standards Organization developed a standard called OSI (Open System Interconnection), which defined seven different layers of protocols. Each layer was responsible for a different function and independent of the layers above and below.

For example, the most popular protocol today is IP, the Internet Protocol. IP devices have to conform to strict rules, but they don't care about what the IP traffic is actually used for or what it travels over. Below IP lies the physical network infrastructure: a LAN, a telephone line, a fiber-optic cable, or a radio link. Above it are the applications that run on the Internet, such as Web browsers and email.

No real protocols follow the OSI model exactly, but networking professionals frequently refer to them, so they are worth knowing about. The seven layers, along with an added security layer, are:

  1. Physical Layer. This layer covers the actual transmission medium, such as radio waves. It also governs the type of modulation used (amplitude, frequency, or phase) and the design of interface plugs.

  2. Data Link Layer. In a fixed network, this layer corresponds to the type of LAN architecture, usually Ethernet. In a wireless network, it covers the TDMA or CDMA multiplexing schemes of standards such as GSM and cdmaOne.

  3. Network Layer. This layer is responsible for actually transferring data between different machines. The most common is IP, which routes small packets or datagrams based on an address encoded into them. Others include IBM's NetBui and Novell's IPX (Internet Packet Exchange).

    1. Transport Layer. This layer covers error control and prioritizing traffic. On the Internet, the original Layer 4 Protocol is TCP (Transport Control Protocol), which is connection-oriented like USSD; it waits until an acknowledgement that one datagram has been received before sending another.

      TCP is still used for fixed Web surfing and email, but an alternative, called UDP (User Datagram Protocol), is becoming more popular for wireless access and streaming audio or video. This is connectionless like SMS, and so just sends datagrams off into the ether without checking that they arrive. This makes it less accurate, but faster.

      WAP can use UPD/IP, but as an alternative also contains its own protocol called WDP (Wireless Datagram Protocol). This covers both Layers 3 and 4, resulting in more efficient transmission.

    2. Security Layer. This layer is not part of the OSI spec, but both the Web and WAP include optional encryption standards that sit just above the Transport layer. The wired Internet's standard is called SSL (Secure Sockets Layer) and is used when connected to Web sites that ask for sensitive information, such as credit card details. Many browsers use a key symbol at the bottom of the screen to show when it is in use.

      WTLS (Wireless Transport Layer Security) did not actually appear until version 1.2 of the WAP standard, published at the end of 1999. This meant that the first generation of WAP phones lacked any security features except GSM's own encryption, and is why the acronym is sometimes pronounced witless.

  4. Session Layer. According to the OSI standard, this is supposed to control who can send or receive information. However, on the wired Internet it is irrelevant. SSL sits where the Session Layer should be.

    As if to make up for this, WAP requires two protocols at this layer. The lower one, WTP (Wireless Transaction Protocol), is designed to ensure that data has actually been sent correctly; it checks for error messages from the recipient, then resends the data if necessary. The wired Web doesn't need this, as it incorporates error control into TCP at the Transport Layer.

    Above WTP is WSP (Wireless Session Protocol), sometimes pronounced whisper. It does nothing itself, but enables traffic to bypass WTP or WTLS if error control or encryption are not needed.

  5. Presentation Layer. This is supposed to control how data is presented to applications, but it is very rarely used.

  6. Application Layer. This is the one most familiar to users and can often be seen in Internet addresses; the first few letters tell the browser which protocol to use. The most common on the Web is HTTP (Hypertext Transfer Protocol), but FTP (File Transfer Protocol) is also popular. Email has its own protocols, SMTP (Simple Mail Transfer Protocol) and POP (Post Office Protocol), as do individual applications, such as the distributed music system Napster.

    In WAP, the Application Layer is known as the WAE (Wireless Application Environment). It supports three different applications:

  • WML (Wireless Markup Language), the XML-based variant of Unwired Planet's HDML

  • WMLScript (or Wscript), a scripting language based on Java-Script.

  • WTA (Wireless Telephony Application), an interface which allows WAP to access the phone's features, such as the address book, and even to make calls. Like WTLS, this was not added to the standard until version 1.2, so was not implemented properly in the first batch of WAP phones, which appeared in early 2000.

Bearers

WAP is designed to be bearer-independent, meaning that it can run over any wireless technology. It is really designed for digital cellular systems, but also works over shorter-range radios. For example, Finnish telecommunications company Sonera has actually demonstrated how WAP and the single-chip Bluetooth radio can operate together, by bypassing the cellphone network when two devices are close together. It is testing Coke machines equipped with a WAP server and a Bluetooth transceiver, enabling any users nearby to access a menu of drinks. If they too have a Bluetooth chip, the menu appears in their WAP microbrowser, and the costs of any drinks they buy are added to their phone bill.

Most of the first real WAP services ran over GSM, usually requiring the user to dial a special phone number as they would when calling an ISP (Internet Service Provider) from a fixed phone. This is inefficient for two reasons. It means both that the user's phone line is tied up and that they are charged for every second that they remain online. For most of this time, the connection is being wasted—the phone transmits or receives only for a fraction of the time.

Bursty traffic is ideal for packet-switching, which does not require a phone line to be kept open. Some operators have experimented with running WAP over SMS or USSD, allowing the user to make phone calls and surf the Web simultaneously, but the small limits of the messaging services make them too slow. Many analysts think that WAP will really take off when it is integrated with GPRS, the upgrade to GSM that adds packet-switching and higher data rates. It also works well, though not as fast, with existing packet-switched services such as CDPD and Mobitex.

Figure 6.1 shows the complete WAP protocol stack, along with some of the lower bearers over which it can travel.

Figure 6.1. WAP protocol stack


Architecture

Although WAP uses its own protocols, it is designed to be compatible with the Internet. Pages written in WML can travel across the Internet using regular HTTP over TCP/IP, then be converted to WAP at the gateway between the Internet and the wireless network.

The protocol conversion leads to a security weakness, because the two protocol stacks use different encryption systems. Data is encrypted over the Internet using SSL; it's also encrypted over the wireless link using WTLS. But in between, it is vulnerable. Operators of sites where security is important must either act as an ISP or fully trust whomever their customers are dialing through. In most cases this is the mobile operator, though it could be anyone.

WAP is often described as a network-centric protocol, meaning most intelligence is embedded in the network rather than the comparatively "dumb" phone. This can cause compatibility problems, particularly where WMLScript is concerned. While regular Internet browsers can contain an interpreter for their scripting language, most mobile phones can't. Instead, the script has to be compiled by the gateway, a process which depends on the type of processor chip inside the phone. To ensure that WAP actually works, gateways have to be tested with every processor, then upgraded whenever a phone based on a new processor is released. This problem is exacerbated by the lack of standards for phone processors; while the Intel x86 acts as a de facto standard for PCs, there is as yet nothing comparable in the cellphone world.

WWW:MMM

WAP phones often don't look any different from their non-Internet enabled counterparts. They may have a slightly larger screen or a wheel to navigate around it, but even these are no guarantee. Many phones have begun to include screens to play simple games or display SMS messages.

To make WAP devices easier to identify, most are marked WWW:MMM, which stands for World Wide Web: Mobile Media Mode. The phone companies hope that this will become a cross-industry brand, eventually lasting beyond the lifetime of WAP.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
18.117.11.247