Which Wireless? Choosing Wireless Devices for Infrastructure Implementation
A system designer tasked with implementing wireless support for a mobile workforce today is faced with a bewildering variety of choices. The designer must decide which technologies support their application, and then the designer must reconcile these technologies with carrier availability, equipment variations, and user preferences. An overview of the current state of the art and a frank look at the options available can help us decide what sort of wireless applications can be implemented.
The primary question is probably one of coverage. Many technologies are limited in the areas they cover. The next question is speed or bandwidth. Higher speed means that more data can be delivered. In practice this means that with high speed we may deliver multimedia content or pictures. Medium speed is good for large amounts of text and small pictures. Slow speed or low bandwidth is good for short text messages. Another important consideration is latency, or how long it takes to deliver a message. Paging systems can have delays of 30 seconds to several minutes. Some latency may be introduced by the underlying technology. Web pages or wml sites may introduce delays when they are under heavy usage. Two other considerations are building pentration and if the communication is a two-way system or a one-way system.
Some technologies are almost everywhere. Paging technologies are able to get a message to almost anywhere in the continental United States. First generation analog cellular telephony also provides national coverage with few “dead areas”. Cellular phones are currently being used with paging technology for field communications. With the addtion of computer telephony integration (CTI) it is possible to achieve a limited degree of enterprise integration. Solutions of this nature are currently common and available with off the shelf technology, although integration to enterprize systems may present some challenges.
This type of solution limits the amount of interaction that a user can have with the enterprise because the users are limited to responding with numeric entry from the phone’s keypad. The most usual response to this is to limit input to selections from menus and numeric entry only. Some attempts have been made to integrate voice recognition systems with this methodology, but the current state of speaker independent voice recognition systems is not always up to the challenge. In addition, integrating speech recognition with other packages is often a daunting task. These systems tend to just use the pager to tell someone when to call a number with the cell phone.
The next technology to examine is two way paging. Two way paging vendors have about 70% of the population of the continental US covered. This coverage tends to be best near urban areas and major highways, but is somewhat better than digital phone coverage. It compares favorably with analog phone coverage. The first upgrade to our system, assuming that we need national deployment, should be to include digital two way paging. There are two primary systems for two way paging. The Research In Motion (RIM) pagers use a system called Mobitex, which has a somewhat restricted deployment. Motorola and other vendors use a system called ReFlex that has a somewhat larger coverage area. There are also other competing systems. Many of these systems can be integrated with email systems, but depending on paging carriers the message length may be fairly short. The most common size is about 500 characters.
This type of system has an improved user interface because most of these devices have small keyboards. It’s probably not a good idea to have large amounts of text entry, but the limitation is more one of message size limitiations than entry problems. This system does have a high latency though, so it should be used more for sending in forms and receiving short information and notification messages.
We have to limit our deployment somewhat to achieve an improvement in functionality. Digital telephony requires either GSM-NA (a system similar to the system in Europe, but operating at 1900 mhz) or CDMA technology. This technology is available in most urban areas and near major highways. It is important to spend time with coverage maps if you are building infrastructure that will leverage smart phones to verify that the features you need are implemented in your target areas. Urban areas usually provide one or more of the technologies required. You will probably need to use more than one type of mobile terminal (smart phone) to achieve national coverage. Since each type of smart phone has slightly different features it will be necessary to customize based on available services. This may lead to a “lowest common denominator” approach. These systems often do not meet the needs of the end users.
The smart phones have several interesting features. The first of these is the somewhat inappriately named wireless web technology. This allows the user to view documents formatted in wireless markup language (WML). Another interesting feature that may be available is short message service (SMS). These two technologies work well together when both are available. SMS allows a short alpha-numeric message to be sent to the phone. In a properly functioning system this turns the phone into an alpha-numeric paging receiver. Some systems do not handle this properly though. On some devices you must be in a “data mode” to receive SMS messages. This has the effect of causing you to not be signaled when messages come in, but only when you go to the data mode.
These devices are currently limited in data bandwidth as well, generally having bandwidth in the 9600 to 19200 bit per second (bps) range. The current systems deal with this by splitting the messages into small packets and using data compression techniques to reduce the number of bits sent at a time. Given the small sizes of the screens on these devices this is acceptable for most users. System design should take this into account, so that the user isn’t constantly asking for more data.
Wireless modems and other wireless connectivity devices use the same data handling methods. We can use smart phones and radio modules (mostly CDPD modules and CDMA phones) as modems over the serial cables supplied with most handheld devices. These provide low speed data links (9600 to 19200 baud) links. This type of coverage works a lot like the wired modems of a very few years ago. This type of connection works acceptably well when being used to transfer data from one application to another, but is not well suited to the kinds of browsing activities made popular by the World Wide Web and other internet applications. Wireless markup language (WML) and wireless application protocol (WAP) were introduced to mitigate this problem. These technologies are commonly available but are not well accepted to date.
Robust applications are possible using this technlogy with “thick” clients. These clients store much of their graphical content locally or generate it from the data transmitted. Thick clients require much more management and maintenance than the newer sort of web-based thin client applications. This sort of solution may be attractive for organizations with special needs, such as sales forces or mobile point of sale systems.
Bluetooth transceivers and 802.11 devices provide much higher bandwidth connections, but require networked base stations and have limited range. These work well when you’re supporting mobile workers who are mobile within your own site. In some areas, there are broadband wireless providers such as Riccochett who can provide 128kbps service. This compares favorably with a wired internet connection. Unfortunately this type of service is only available in a few metropolitan areas at present. It is expected that within the next few years third generation cellular data services will allow broadband connections to be built out into more areas, so if you’re planning a multiple year phased deployment the risk may be worthwhile. Be warned though, you will be at the mercy of outside vendors should you need to deploy into specific market areas.
Looking at the technology constraints we see that we are constrained to either specific cities or providing limited range base stations for current broadband applications. Applications that would stream audio or video, or applications that move a lot of data such as high frequency telemetry may not be achievable without specialized infrastructure. This should improve soon, but must be considered to have higher risk in the next year. Generally speaking we can have high bandwidth over a short distance, or lower bandwidth over a larger area. This limitation will ease over the next few years, so it is essential that we plan our systems for growth.
The following table did not appear in the article, but may be useful when comparing technologies.
|
Technology |
Coverage |
Speed (Bandwidth) |
Latency |
Building Penetration |
Two-way communication |
|
One-way
pager |
National |
Low |
High |
Excellent |
No |
|
Two-way
pager |
National |
Low |
High |
Good |
Yes |
|
WML on
Phone |
Metro/ Highways |
Medium |
Medium |
Fair |
Yes |
|
SMS on
Phone |
Metro/ Highways |
Medium |
Medium |
Fair |
No |
|
CDPD
Modem |
Metro/ Highways |
Medium |
Low |
Fair |
Yes |
|
Circuit
Switched Data |
Metro/ Highways |
Medium |
Low |
Fair |
Yes |
|
Broadband
(Ricochett) |
11 cities |
High |
Low |
Fair |
Yes |
|
Bluetooth |
Campus/
Building |
High |
Low |
Poor |
Yes |
|
802.11
Network |
Campus/
Building |
High |
Low |
Fair |
Yes |
Please tell me what you think about this content and how I might improve it.