Making The (Up)grade
By Fred Dawson
As 1999 comes to a close, operators can for the first time tackle their expansion agendas with an arsenal of base station options that largely address the concerns that until now have made future-proofing the network an extremely hazardous proposition. As a result, the prospects for global penetration of data-ready CDMA-based platforms, including not just cdmaOne and CDMA2000 but also Wideband CDMA as a 3G solution to GSM migration, appear much brighter than they did a year ago. "Interest in IS-95 (cdmaOne) is heating up around the world," says Mark Whitton, director of CDMA access product management at Nortel Networks. "We're not just seeing a lot more capital going into building out existing networks; we're seeing major new-build choices swinging to IS-95."
In part, the acceptance of CDMA throughout much of Asia and Latin America as well as the U.S. is a result of the considerable progress made in establishing a 3G migration path for implementation of capacity expansion which accommodates data as well as voice services, including the promise of a reasonable degree of interoperability between the two main versions of third-generation CDMA platform, CDMA2000 and WCDMA. Equally important, though not so widely publicized, vendors have come up with a range of cdmaOne base station solutions that accommodate virtually all of the operating community's needs in terms of coverage, space efficiency and functionality. For example, Telstra is now testing what Nortel calls the "Boomer" macrocell BTS (base transceiver station), a product which provides coverage from an IS-95 platform over a distance of 180 km and which Nortel plans to introduce market-wide in the second quarter of 2000. While the system is no substitute for coverage by multiple base stations in urban areas or other places where capacity or line-of-sight issues prevail, it is ideal for accommodating rural build-out needs in places such as the midwestern and southwestern regions of the U.S. and in Australia, where the landscape is relatively flat.
Along with the development of base station solutions that accommodate a multitude of build-out requirements, vendors are taking action that will make it much easier to mix network elements from multiple vendors in any given service territory. This point is especially important to the evolutionary potential of CDMA, notes Wendy Fulk, vice president of marketing and communications for CDMA systems at Ericsson. "I get the sense that people are moving very aggressively to expand coverage, and the opportunity to use multiple vendors is an important part of that," she says. Fulk credits Sprint PCS with spearheading efforts to tighten the IS-634 specification that allows a BSC (base station controller) from one vendor to connect with the MSC (mobile switching center) of another.
Another key network performance development progressing toward commercialization is the better performance of "soft handoffs" in cdmaOne systems. This, too, has been a matter of great concern at Sprint PCS, according to Michael Robinson, vice president and general manager of network services.
Now that network operators are using multi-carrier BTSs-base stations that support two, three or even four 1.25MHz RF channels per frame module-they have to install "transition sites" or extra base stations to accommodate the cross-frequency stepping that allows a soft handoff in the multi-carrier environment. This is still an improvement over the one-carrier-per-frame environment of earlier generation base stations, says Robinson, because with more carriers per frame the base station is supporting a higher number of users on a given module, thereby lowering the costs of expansion.
However, ideally, the need for a transition site would be eliminated, thereby maximizing the cost-efficiency gains in capacity while maintaining the soft handoff capabilities of single-carrier frames. "We've been testing use of pilot beacons as one approach to solving the problem, and the results are promising," says Robinson. Pilot beacons, operating off the local geopositioning system (GPS), are used to track users' signals between adjacent base stations, telling the system which base station a mobile is communicating with at any moment. This information allows the base stations to make the stepping transition for the mobile from one carrier frequency to another without the assistance of a transition station.
As well as supporting higher capacities per frame, the new generation of cdmaOne base stations is designed to be easily upgradeable with a circuit pack changeout to the first level of the CDMA2000 G3 specification, known as 1xRTT (Radio Transmission Technology). This is possible because, once there are three 1.25MHz IS-95 carriers operating in the base station-to-RF frame, the move to the 5MHz carrier requirement of 3G can be accomplished with a changeout of just the digital components, leaving the RF components alone. Says Cindy Christy, vice president for AMPS and PCS product management at Lucent Technologies, Inc: "Customers want us to deliver solutions that leverage existing infrastructure to the maximum extent possible."
While the details of harmonizing 3G systems remain to be worked out, vendors say 1xRTT specifications are solid enough to permit the development of chips that will allow testing of the platform to begin next year. According to Robinson, Sprint PCS, which recently introduced a 14.4kbit/s circuit-based data service, remains confident it will be able to bring 1xRTT services to market by late 2000 or early 2001.
As the first level 3G implementation, 1xRTT supports a 144kbit/s data channel, doubles the voice capacity per carrier and sets the stage for the jump to 3xRTT at data rates of up to 2Mbit/s. Because the change to 5MHz channelization will be accomplished with implementation of 1xRTT in base stations, the move to 3xRTT will be done with just a software upgrade.
In expanding their networks, carriers such as Sprint PCS, now in its third build-out phase, are using the new generation of base stations that will support these simple upgrade steps to the 3G platform, thereby reducing the cost burden of future data service expansion.
IS-95B has not won wide vendor support owing to many carriers' preference for the migration path made possible with the move to 1xRTT. But Motorola, by delivering a commercial IS-95B system, believes it is giving carriers who don't want to wait for the 144kbit/s 1xRTT capability a chance to move quickly to higher data rates, using both older and the newer base stations.
Operators in the U.S. as well as Japan and elsewhere are rethinking the IS-95B option, Campbell adds. "There's definitely an advantage to 1xRTT over IS-95B, not only with respect to higher data speeds and double voice capacity, but also because there is greater efficiency at the MAC (media access control) layer and because it opens the path to 3xRTT," he explains. "But you can move to IS-95B and still be well positioned to go to 3G, and people are looking at doing that, especially in places where the demand for data and better data rates is stronger than anticipated."
Motorola has made the data channel over IS-95B a less capacity-consuming monster than most people anticipated, which means that companies such as Japanese cellular operators DDI Corp. and IDO Corp. can respond to rapidly rising demand for improved data services without making major sacrifices in voice coverage. "With the rate allocation algorithms we're using, operators can assign priorities to voice or data and make use of the RF resources accordingly," says Campbell.
This means that RF resources can be assigned to data as need dictates but, if voice takes priority, the capacity is available to accommodate it, with dynamic assignment of carriers according to real-time changes in usage patterns.
Korea Telecom Freetel, the first operator anywhere to implement IS-95B, has been conducting trials with various vendors, including Qualcomm, Inc., which is supplying its MSM3000 chipset and high-speed system software solution. According to Don Schrock, president of Qualcomm CDMA Technologies, KT Freetel's PersNet service will soon allow customers to take advantage of electronic commerce, on-line stock trading, on-line banking and travel reservations via the Internet through handsets that use the MSM3000. The system will deliver data at up to 86.4kbit/s over IS-95B, giving it a 33 percent improvement over data rates set in the standard specification, he adds.
Where the data rate question is concerned, Ericsson, having acquired the base station business of Qualcomm and with it the core technology, is in as good a position as anyone to prepare for the migration to 3G.
"We're taking the approach of looking at both WCDMA and CDMA2000 and trying to avoid doing two separate product designs," Fulk says."We're going to keep our focus on all the elements of the design that can be shared by both groups, so that we minimize the distinctions between the products."
Over the past couple of months this process has revealed many areas of commonality and led to a tightening of the development schedule for Ericsson's 3G products. "When you look at the types of services people are talking about introducing in Japan and Europe-video e-mail, for example-you realize we're not overstating the case for 3G," says Fulk.
Nortel, too, is optimistic that the harmonization issues can be worked out in support of rapid development of a final 3G standard, according to Whitton. "The systems won't be completely converged in the sense that they are exactly identical, but they'll be close enough to support interoperability among various iterations with multimode phones," he says.
Notwithstanding the considerable progress on interoperability across multiple platforms, however, industry leaders realize a deeper level of integration will soon be required if data-based services are really to affect the wireless domain on a wide scale. As Bill Wiberg, president of AMPS/PCS wireless networks at Lucent, notes, without universal agreement on the wireless data infrastructure, carriers are risking a high-cost, low-return segmentation of the market that could make data a non-starter. "The ease-of-use issue has to be worked out, and the way it can be worked out is through partnering among the different members of the industry, be they applications providers, end-terminal providers or infrastructure providers," he says.
To that end, Lucent recently joined with AT&T Wireless Services, BT, Rogers Cantel, Ericsson, Nokia, Nortel, Telenor AS and Telecom Italia Mobile to promote the creation of an Internet Protocol (IP)-based wireless system for 3G systems in the UMTS domain. The new group-3G.IP-wants to see next-generation wireless services such as IP voice, high-speed data and video conferencing based on an architecture that uses an evolved implementation of GPRS (General Packet Radio System). From the GSM perspective, this makes a lot of sense. "We'll see some GSM operators skip to GPRS at 128kbit/s as soon as next year," claims Yankee Group analyst Phil Redman.
However, another, competing initiative is in play. Spearheaded by Cisco Systems, Inc. and Motorola, it purports to be completely agnostic with regard to air interfaces. The migration path to an all-IP infrastructure should start with the creation of packet-based interfaces between the complex operational protocols of the wireless domain and points of interconnection between the backhaul network of the wireless system and the IP "cloud" that is common to all IP networks, says John Shantz, vice president of market development at Cisco. "Traffic efficiency is better than two-to-one on the number of calls you can maintain over a T-1 or other connection using packet technology as opposed to circuit technology," he claims.
The Cisco/Motorola strategy is to offer products based on open architecture for testing immediately, allowing carriers to begin the transition to the IP domain as they launch data services over their existing platforms, regardless of the air interface they use. Eventually, the goal is to accomplish all of the transcoding between the two domains right at the base station, so that all components of the wireless network infrastructure are communicating in the language of IP.
Somewhere in between these two stages, it will be possible to start putting voice signals onto the packet backbone, allowing carriers to avoid the costs of circuit switching as they deploy new infrastructure, according to Shantz. "Our intention is to make use of the same distributed call agent model for wireless IP voice that we're pursuing in the wireline market," he says. "Wireless companies recognize they have to move to an integrated data services platform. This may not be good news for some of their traditional suppliers in the switched-circuit market, but maybe we're seeing a move to a customer-driven rather than a manufacturer-driven marketplace."
The good news for operators is that, as they consider the vast possibilities of data-based services, the equipment now available to support near-term expansion needs supports the long-term migration path they are looking for, no matter what the ultimate IP architecture turns out to be.