Will ?Mobile Broadband Wireless? Redefine Access?
from the March 2000 issue of BCR ACCESS, a supplement to Business Communications Review, pp. 14?19
bcr.com
by Craig Mathias, a principal at Farpoint Group.
Cable modems and xDSL continue to drive the rapid expansion of the Internet-access market, but fixed wireless communications also plays a role. Fixed wireless offers much shorter time-to-install, lower operating costs, greater installation flexibility and performance equal to or greater than many wireline options, so it's not surprising that a broad range of carriers and equipment suppliers now play in the fixed broadband wireless space (see "The Wide-Area Wireless LAN")
More problematic, however, has been the evolution of mobile broadband wireless. Central to this issue, as one might guess, is the mobility factor itself. In fixed wireless configurations, communications between network nodes is almost always line-of-sight, and the relative positions of the nodes do not vary. As a consequence, opportunities for interference and other radio-communications artifacts are minimized. Overall traffic engineering is fairly straightforward, with only a few path-performance calculations to perform.
Furthermore, both licensed and unlicensed products and services are available for fixed wireless, and concern about interference in the case of unlicensed systems is usually overblown. By combining spread-spectrum radio and error-correction techniques, unlicensed products can offer throughput to 100 Mbps and reliability essentially equal to that of licensed products.
But mobile communications systems have no such luck. Mobility introduces a wide range of problems, including:
Omnidirectional antennas, which can transmit or receive from any direction, are needed, because the mobile node doesn't know the exact location of the other end of the connection (normally a base station in a cellular system). Omnidirectional antennas waste energy when transmitting, and allow potentially interfering signals (such as those resulting from multipath) into the radio when receiving.
Mobility itself introduces problems in the form of fading: The loss of signal strength when the receiving antenna is someplace the transmitter's signals have trouble reaching.
Finally, support for mobility requires broad coverage. Current cellular networks have patched together reasonably large areas of coverage, and nationwide roaming is possible in many cases, thanks to inter-carrier agreements or the wide footprints of very large carriers like AT&T and Sprint PCS. Nevertheless, we still live with the familiar indications of no service, system busy and dropped calls.
This is not to say that mobile broadband wireless represents an insurmountable technical challenge?far from it. While the core technology is sophisticated, the primary system elements required to justify investments in development have now emerged. Business and consumer demand (primarily for mobile Internet access), an improved regulatory environment and a solid track record of consumer-grade wireless are all paving the way?even though mobile networking nirvana is still a few years off.
The Here and Now
In the meantime, there is already a wide range of wireless services suitable for remote access to both the Internet and corporate networks?albeit at slow speeds, and with spotty availability. Many services can nonetheless offer reasonable coverage, roaming and affordable prices.
What they don't offer is throughput commensurate with landlines?while 56K modems constitute the floor of most modern communications networks, wireless networks are still stuck in the 9.6/14.4-kbps era. The reasons for this are simple?today's wide-area wireless systems were designed, like the public switched telephone network, with voice in mind. The technologies (and business models) behind today's services were driven more by the need for increasing volumes of low-bandwidth, isochronous, full-duplex voice.
What we thus have today is a set of services which are fairly well-suited to messaging and vertical applications (often with relatively high latency), but not to contemporary datacom or access to the Web (Figure 1). While Web access and data communications are possible over today's wireless networks, at least to a limited degree, their use can be painful in practice.
For example, dialup access via cell phones with appropriate modems can work, but is often slow and relatively expensive. And circuit data connections on digital cellular, while defined in all three major digital standards (GSM, TDMA and CDMA) are not widely available from the carriers. Paging networks and the short-message services (SMS) offered by many digital cellular operators, although well-suited to messaging, can have unacceptably high latency for more sophisticated applications. Packet radio networks (American Mobile's ARDIS, BellSouth Wireless Data, and the CDPD services offered by many cellular carriers) offer less latency and a greater orientation towards data (as opposed to messaging) applications, but are still limited to the 9.6-kbps range?and often less, depending upon traffic, range and prevailing radio conditions.
But we have seen some exciting additions on the subscriber unit front, including 3Com's Palm VII and Research in Motion's (RIM) Blackberry. The Palm VII adds BellSouth Wireless Data service to the most popular organizer platform, and the Blackberry includes a small but usable keyboard on a two-way email/organizer device the size of a pager.
Half Way There
The big news today in wide-area wireless data is the advent of what are being called "2.5G" systems. The first generation ("1G") of cellular was the analog systems that are rapidly falling into disuse in many countries?Australia, for example, turned off much of its AMPS service this past New Year's Eve. Second-generation cellular includes digital services based on GSM, TDMA, and CDMA.
It's important to note that carriers "went digital" more for reasons of capacity than to gain additional features, but each digital standard included limited support for data. With the market for voice exhibiting some signs of maturity (decreasing per-unit utilization and continually falling prices), carriers are beginning to look to data for both revenue growth and profitability.
2.5G is a stepping stone on the way to true third-generation (3G) functionality, which will more than equal today's landline services. Essentially, 2.5G builds on the core functionality in existing wireless standards without major changes in infrastructure:
The latest update to the CDMA standard, IS-95B, includes support for data at 64-115.2 kbps. Some carriers will deploy IS-95B in 2000.
GSM systems can take advantage of the General Packet Radio Service (GPRS), which can offer up to 57.6 kbps in each direction, and its circuit-data counterpart, High-Speed Circuit-Switched Data (HSCSD), with 64 kbps (or even greater, depending upon implementation). These services are already being tested in Europe, but it's unclear when or even if North American operators using GSM on PCS frequencies will adopt them. The reason? Business issues?while a migration to GPRS and HSCSD could expand the overall opportunity and make "better" use of bandwidth, it also takes bandwidth away from the current and profitable traditional uses, and it introduces new levels of technical complexity which in turn spawn a whole host of requirements that relate to training, customer support, marketing and hardware/systems-integration. As a result, carriers won't move in lock-step toward these capabilities.
The TDMA standard, IS-136, receives an upgrade in the form of IS-136+, which uses improved modulation to yield data speeds as high as 64 kbps. It's not clear, however, if this technology will be deployed?as noted below, the TDMA and GSM camps are merging to a great degree, but GSM is incompatible with IS-136.
There are other opportunities. Metricom, for example, is deploying its base stations on light poles around the country, thanks to $600 million in new investment. The Metricom Ricochet service is essentially a mobile connection to the Internet, and speeds in Metricom's next-generation product will reach 128 kbps. And Qualcomm's IS-95 HDR product can provide 1.2?2.4 Mbps on a standard 1.25-MHz IS-95 channel.
Third Base
3G systems complete the migration of cellular to a true broadband service. The vision of 3G is simple?anything landlines can do, wireless can do as well. 3G cellular will include voice (with improved quality), messaging, location tracking, data (circuit and packet) and better battery life.
Perhaps the most important specification is peak throughput of 2 Mbps when stationary, 384 kbps when moving relatively slowly and 144 kbps at high (automobile) speeds. Note, again, these are peak throughput numbers, and represent traffic only from the base station to the mobile unit, as the channel is asymmetric. Nonetheless, this level of performance would eliminate the gap between fixed and wireless in all but a few applications.
On the other hand, we have yet another set of alphabet-soup acronyms to deal with, and little hope of a single uniform technology across a single uniform set of frequencies worldwide. As with 2.5G, each major cellular technology will evolve toward 3G along its own path (Figure 2); production systems might appear as early as 2001 in some parts of the world (most notably Japan), and by 2003?2005 for the rest of us.
By far the most important technology in the 3G story is wideband CDMA (W-CDMA), wherein CDMA techniques are used in radio channels of 5 MHz or more. More bandwidth means more users, or more capacity for data, or both. Since CDMA can be implemented in a variety of ways, several warring camps have been lobbying hard to get their version blessed as "official."
The International Telecommunication Union (ITU), which is working to establish global recommendations on 3G technology via its IMT-2000 project (IMT= International Mobile Telecommunications), has basically given up on trying to have a single global air interface; instead, the ITU issued five not-yet-official recommendations. Of these, there will likely be at least three or four major implementation contenders (discussed below), each with its own backing from powerful trade groups. And we can't rule out other technology entrants over time.
The TDMA-based major players, GSM and IS-136, will eventually adopt WCDMA (a specific implementation of W-CDMA) as their third-generation technology, in the form of the IMT-DS (International Mobile Telecommunications Direct Sequence) recommendation. It's possible in some cases to operate W-CDMA systems as "overlays" on TDMA spectrum, easing deployments for many carriers?3G doesn't necessarily render 2G and 2.5G systems obsolete.
There is even a place for Digital Enhanced Cordless Telecommunications (DECT), a cordless telephone technology standard popular in Europe. DECT has roughly 1.1 Mbps of data capacity per channel; presumably it would be deployed with closely-spaced cells. There is also a proposed CDMA-based ITU recommendation (IMT-SC) that can operate on unpaired radio channels using time-division duplexing.
The lack of inter-system client interoperability implied by the above is perhaps to be expected, given the rich history of cellular, global politics, and the established presence of several technologies. While broadband data services will be available in many parts of the world over the next five years, global roaming remains an issue that still needs to be resolved.
Home Run?
Of course, there are still a lot of details to iron out as wide-area wireless goes broadband. Among the key challenges are:
Bandwidth availability: Spectrum has not been allocated in the U.S. for 3G at this point, and it's very unlikely we'll see global commonality in spectrum allocations. While this doesn't preclude development of a "world phone," it does push its availability beyond most people's planning horizons.
Deployment: Buildouts take a long time, with all of the usual hassles regarding siting, real estate, network configuration and NIMBYs. The wide availability of 3G (i.e., on the order of today's PCS deployments) is at least four years away for most potential users.
Markets: While there is much excitement over the possibilities inherent in mobile broadband, this interest stems largely from wireless equipment manufacturers, who see a gold mine in the new technology. It is not clear that the wireless carriers and operators are quite as excited. Indeed, they are more likely concerned about possible disturbances in their cash flow and in the competitive balance?churn (turnover in the customer base) has always been a problem in the cellular industry. And finally, it's not at all clear that consumers will flock to mobile broadband; they are just getting used to the low-bandwidth wireless Internet services , and we still lack such crucial details as pricing for mobile broadband services.
Regardless, there is cause for optimism. The technologies are there, the interest on the part of equipment manufacturers and at least some carriers is there and, presumably, the demand from businesses and consumers will materialize. After all, voice communications and messaging are no longer sufficient for businesses and consumers. The Internet?and through it, access to corporate networks?is now just as important. |
