The following is an excerpt from the August 1998 Wireless Computing Mid-year Update, from Andrew Seybold's Outlook newsletter.  Andrew Seybold is a leading and oft-quoted industry analyst within the microcomputer and communications fields. He heads a consulting firm specializing in mobile computing and advanced communications technology.  

Below is Andy Seybold's view of today's wireless data opportunities in the United States.

The following does not reflect the views of Intel Corporation and does not constitute an endorsement.

Circuit-Switched or Dial-Up Data Access

Key to the success of any new technology is how easy it is to deploy, and whether the perceived value outweighs the cost. Before cellular phones were introduced in the early 80s, mobile phones were available in many parts of the U.S. Mobile phone services were run by phone companies and Radio Common Carriers (RCCs), and they provided either manual or automatic connection to the wired phone system.

Only those who really needed a phone in their car had one. In Los Angeles, where I was living at the time, users numbered in the thousands and the available channels were congested. In some cases, we had to wait up to thirty minutes for a dial tone. When we were able to make a connection, we paid several dollars a minute for the privilege of talking on the phone while we were driving down the freeway.

The development of cellular phones solved many of the problems associated with mobile phones, but it was several years before the number of users approached what was needed for such systems to be profitable. During the first five years of cellular build-out, if you had a cellular phone you were considered "rich" by most standards. Car phones sold for thousands of dollars and airtime was $0.50 to $0.75 per minute, depending on the time of day or location of the system. (In California, where we have a Public Utility Commission that is watching out for our best interests, we pay some of the highest rates in the nation!)

Coverage was not great in the early years. Most systems provided good coverage along major freeways, and gradually added business areas over time. It was many, long years before networks were built out to provide coverage in the suburbs, and longer still before they reached some of the more remote areas.

One of the stated reasons that the FCC auctioned Personal Communications Services (PCS) frequencies was to increase competition, thereby driving down the per-minute cost of using cellular phones. But even before PCS, cellular carriers had begun reducing per-minute usage costs and offering pricing bundles that included "free" off-peak minutes of use.

The real growth in cellular phone usage began in the late ’80s and early’90s. By about 1993, the pace picked up considerably. This was primarily due to lower prices and the introduction of a number of handheld phones (as opposed to car-mounted phones). Cellular phones had finally become small enough that people considered carrying them. A number of other factors were behind the increased adoption rate of cellular phones:

• Companies adopted voice mail for use by their employees. Accessing voice mail and being able to return calls during "dead" travel time became a productivity gain both for the employee and the company.
• Pricing continued to drop, and some phones were available for "free" when signing up for service.
• Cellular phones work just like wired phones except that you have to remember to hit the "send" and "end" buttons.
• Having a cellular phone enables you to be connected to anyone with a phone, even while you are in your car, riding the subway, or walking down the street.
• The Silicon Valley culture and the advent of the personal computer require that we all run faster to keep up. Fax machines, instant accessibility, FedEx, turning on a dime, reinventing business, downsizing, and all of the other business-related issues keep pushing us to accomplish more with fewer resources and less time.

 

Today’s Mobile Work Force

Today there are approximately 45 million mobile workers in the U.S. This number is made up of people who are away from their office 20% of the time or more. Many carry a cellular phone (there are 60 million wireless phones in use in the U.S.). Many carry pagers (there are 40 million pagers). And many carry a palmtop, handheld, or notebook PC (20 million).

Why isn’t this mobile work force enabled with wireless data access? Start with the points listed above and add:/p>

· It is too difficult to connect to the information sources we want and need.

However…

One group of mobile workers could be instantly drawn into the wireless data pool if only the industry would recognize them, put together shrink-wrapped solutions for them, and properly market such solutions. This group is made up of users who already have a mobile computer and already connect via a wired modem to their own corporate resources. There are millions of these folks out there! They have learned by trial and error how to plug in a modem card, attach a phone cord, and plug it into an RJ-11 jack.

They have already struggled with modem software, cursed at set-up strings, yelled at the idiocy of having to open multiple windows to change dial-up access numbers and outbound dial numbers. Yes, they have weathered all of this and have become proficient at using a computer connected to a modem connected to a telephone line. Their IS managers have learned along with them. They have put in modem pools and, more recently, RAS servers so these folks can dial into the corporate data pool. IT managers have figured out how to coax their legacy programs to run over phone lines, provide remote access to their corporate information assets, and support dial-up mobile folks.

The wireless industry has been so focused on packet-based solutions that we have overlooked this group of users. While some of these people may have tried analog dial-up over cellular, most have found that this approach is too difficult to be a satisfactory solution. No one has made a move to win over these folks with dial-up circuit-switched wireless—at least in the U.S.

Today, dial-up wireless services are available over analog cellular or GSM PCS networks. CDMA PCS networks will be able to offer dial-up by the end of the year. The Nextel system is capable of dial-up circuit-switched although it has not been implemented. In fact, the TDMA PCS (D-AMPs) systems run by AT&T Wireless, SBC, and BellSouth are capable of dial-up data as well. For some reason, these companies appear to be bypassing dial-up and heading directly to packet-data systems.

Installed Base

Recapping, we have an installed base of users who dial into their own corporate computing assets a couple of times a day, and some who dial up the web and retrieve information. We have at least one digital network that provides digital dial-up capabilities today, two more coming online this year, and another that could offer the service if it wanted. If this is true, why isn’t there a mad rush of people buying dial-up wireless modems?

The answer is complex--as are all answers in the wireless world--but manageable. Let’s consider the main issues one at a time.

GSM is the only digital standard that supports dial-up today. However, GSM systems do not provide nationwide coverage--Dallas and Chicago are two big holes. GSM in-building coverage is not very good yet, and there are roaming charges outside your prime provider’s area. Moreover, GSM carriers are not interested in data.

GSM PCS networks that are up and running today cover a large portion of the U.S., and they support 9.6-Kbps dial-up data. Before you say that this is not fast enough, try it. I have been using GSM for data for some time now and have had occasion to download from some highly intensive graphic web sites, though I wouldn’t want to make a habit of it. I have dialed into the Internet, connected to my POP3 server, and sent and received my email. I have also dialed into a server on which my calendar and personal information manager (PIM) reside and have synchronized my notebook’s copy of Outlook with the information stored on the server. If I had a RAS server installed at my office, I could dial in and access my network. (We are in the process of setting this up.)

In short, at 9.6 Kbps, I can do everything wirelessly that I can do plugged in from a hotel room, only a little slower. My experience from hotel rooms all over the nation is that I can connect at between 21 and 24 Kbps—not at the 33 Kbps or higher speeds that most people think are available. Wirelessly, with GSM PCS I am operating at about 40% of the speed I achieve from a hotel room. The difference is that I might be sitting in my car driving between appointments, at an airport gate, or anywhere there is coverage but no RJ-11 jack.

What Did I Do Differently?

What did it take to make this work? First, I had to call my GSM carrier and request a second number for my GSM phone. (The GSM system has different phone numbers for voice, data, and fax.) Then I inserted a Motorola PC Card modem and installed software for my StarTac phone to enable it to make a data link. The modem was recognized instantly, and switching back and forth from my dial-up to my GSM modem is a matter of choosing the appropriate modem from the dialer selection box. I could leave both modems in the computer, but since PC Cards can drain the computer’s battery, I am in the habit of taking them out.

I took the portable computer out to the car and drove into Boulder Creek. (While there is no GSM coverage where I live, there is a single cell in town about six miles away.) I turned on the computer, plugged the phone into the PC Card, and instructed the computer to dial. The computer dialed the phone and the next thing I knew, I was connected to the Internet. Total installation and de-bugging time? About twenty minutes!

Differences

It takes a little longer to establish a connection over the GSM network than it does for wired dial-up. Since the connection is being made from a car or other mobile location and you are usually watching it, it seems to take unusually long. However, I timed it on several occasions and it actually takes only 10 or 15 seconds longer. Once connected, the system works the same as any other dial-up connection. It may seem somewhat slower, but when I went to the Net and turned off the graphics, I was quite pleased with the results.

The biggest difference between analog cellular dial-up and GSM dial-up is in the robustness of the connection. I have driven along freeways while connected to a web site or while updating my calendar on many occasions. Not once has GSM dropped a connection. The same cannot be said for analog cellular.

The Cost

The costs are the same for dial-up data as they are for wireless voice, with the caveat that some set-up time is required to make the connection. The modems at both ends must synchronize with each other no matter the method of connection, and it takes airtime. The other issue, of course, is how much it will cost to send and receive information using GSM dial-up data. You might think that at 9.6 Kbps it could cost you a lot of money.

With PCS GSM, most of the service plans include "free" minutes. My plan, for example, costs $50 per month for 400 minutes of airtime. While PacBell does not charge an extra monthly fee for using its Data/Fax services, data calls are not included in the plan’s airtime. Data and fax calls are billed at the normal per-minute rate of $ 0.25. Thus, a typical email session might cost me between $1.50 and $2.00. Of course, when I place a phone call with a wired modem from a hotel room to check my email, I incur surcharges—usually $0.75 or more per call.

PacBell also offers short-messaging services for a flat rate of $9.95 per month with operator-assisted messages billed at $0.50 each.

Comments and Conclusions

I can understand why many folks who are promoting wireless data have not spent much time looking at dial-up connections when analog dial-up is so difficult to implement. And GSM is such a new system in the U.S. that there is a lot to be learned. I can even understand why the European experience with dial-up GSM has not been translated to the U.S. since the European mobile computing community is much smaller.

I think that it is now time for the industry to turn to dial-up wireless as a viable option. In fact, I think that getting users up and running with dial-up wireless is a great precursor for introducing them to the packet-based solutions that the industry believes hold the most potential.

Dial-up systems can move larger files, handle faxes, and otherwise dispense data files that are large and not well suited for packet systems. Moreover, the cost for a minute of airtime has come down dramatically. At $50 per month for 400 minutes of airtime to and from anywhere in the California/Nevada area, I am paying $0.08 per minute. If the time it takes to dial-in, send and receive my email, and disconnect is five minutes (which is a longer than it usually takes me), each call costs $0.40. Not bad when you consider that I can download fifty email messages and send back a dozen or so during the one call!

When CDMA systems come online later this year, they will have several advantages over GSM systems. First, the CDMA dial-up data rate will be 14.4 Kbps. Second, all CDMA phones have a built-in modem. (Only a cable between the phone and a serial port will be required). Further, Sprint’s "free" minutes include all long distance charges in the pricing package we have chosen. Qualcomm is working on a "fast connect’ sequence that it claims will shorten the time between dialing and connecting by about 50%--a time and cost savings that will be appreciated by all.

Circuit-Switched in the U.S.

By the start of 1999, we will have two types of PCS systems that will permit dial-up data access--GSM and CDMA. It is not known whether the TDMA systems or Nextel will make dial-up available. We suspect that if this form of data connection catches on in the horizontal space, both of these carrier groups will be forced to offer dial-up data to stay competitive.

It is interesting that most hardware and network vendors seem to be ignoring this opportunity. Perhaps it is because data speeds are perceived by many to be too slow, or because of a lack of understanding about how similar wireless dial-up can be to wired dial up. Or perhaps it is because analog dial-up has been such a negative experience.

While there is no way to know exactly how many cellular users have tried to use, or are using, analog dial-up, guesstimates put the figure at about 5% of the installed base. If this is true, almost three million wireless users have tried dial-up. If these users discover that dial-up with PCS works far better than with analog cellular, and decide to switch their service to a GSM or CDMA carrier, it could be a significant windfall for network providers who are interested in obtaining and keeping new customers.

It is easy to demonstrate how dial-up works. The set-up as described above is easy. And those who use modems already know how to make it work. Those who spend their online time surfing the web and looking at graphics-intensive web sites should forget about dial-up wireless, as should those who need to send megabyte PowerPoint files. However, there may be times when it doesn’t matter how much it costs to move a file. If the information is of a critical nature and there isn’t an RJ-11 jack nearby, $10 or $20 to send a file is not a lot of money. Most people do not hesitate to send a FedEx package that costs around $12.

Dial-up circuit-switched wireless connectivity has another advantage for the mobile worker. The IS department has no way of knowing whether the access to the modem pool or RAS server is coming in over a wired or wireless circuit. This means that a road warrior who wants to try wireless access does not need to ask anyone’s permission.

This is, of course, the way all new technologies are introduced in corporate America--they are sneaked in the door. Only when there is a critical mass does the IS department take notice. We think that there is a segment of the mobile population that will be willing to try dial-up over wireless, especially with the new digital systems. If we could get all four of the digital technologies--GSM, CDMA, TDMA, and iDEN-Nextel--interested in promoting dial-up wireless, and convince vendors to bundle PC Card modems and/or cables into a dial-up wireless solution, we could find the networks populated with a significant number of users. Network providers might even begin to fight over this market segment.

The Industry

Still, the wireless industry, at least its hardware vendors, seem to be hung up on packet systems for wireless data delivery. We will solve the back-end connectivity problems associated with packet-data connections over the next few years. In the meantime, dial-up makes sense. We can foresee a group of users who would want both technologies available to them. For short messaging, calendar updating, and similar tasks, they would use packet. For file transfers, faxing, and such, dial-up would be used. And, of course, these users would look for an RJ-11 jack for a wired connection whenever possible.

IBM is one of the very few companies that seems to understand this dial-up scenario. A year or so ago, it launched IBM ToGo, which provides the type of dial-up services we have just discussed to its Intranet service users. Unfortunately, IBM started out with dial-up analog phones and a flat monthly price of $99, which is too high. Maybe with the availability of GSM and CDMA data-capable systems, IBM will take another stab at it.

In the meantime, it would be interesting if hardware companies did some creative bundling. How about a wired modem and a GSM modem? A handset vendor and a computer vendor putting together mobile connectivity packages such as a computer with a PC Card and cable for a specific phone? All of these and more are viable options. It simply takes an understanding that dial-up wireless is an extension of dial-up wired and, now that we have digital systems, it works the same way.

For years, the packet folks have been saying that the problem with wireless data is that it is not like wireless voice. On the other hand, wireless voice is very much like wired voice and users can understand the wireless concept. Well, wireless dial-up data is the same as wired dial-up data only a little slower and a little more expensive. Isn’t it better to have some information where before there was none, even if it is delivered more slowly and at a premium?