The Technical Case For Convergence Of Third Generation Wireless Systems Based On CDMA
--Five Key Technical Principles To Consider--
The wireless industry is actively addressing the evolution of technology to
support third generation IMT-2000 systems, systems that will offer high burst
rate packet communications in addition to high quality voice and medium rate
data. Several radio technology proposals have been put forth and virtually all
are based on direct sequence spread spectrum Code Division Multiple Access
(CDMA) digital wireless technology. As the first in the world to develop this
technology for commercial use, and, as a result, the inventor and patent holder
of the core technologies enabling efficient commercial application, QUALCOMM
welcomes and supports this recognition of the overwhelming advantages of
CDMA over earlier technologies. The present generation of CDMA, denoted as
cdmaOner, has been standardized and is in widespread use for cellular,
personal communications, and wireless local loop in many countries. CDMA is a
proven technology, providing the highest voice quality, and rapidly evolving to
support efficient medium rate data. Many companies including QUALCOMM are
working on a multi-bandwidth evolution of cdmaOne, denoted as cdma2000r
(previously called Wideband cdmaOneT). Others are working on a variation of
CDMA denoted W-CDMA, using core technology from cdmaOne but with a
number of different choices for parameters and technical details. QUALCOMM
believes these proposals should be converged to a single, global third
generation standard working equally well with both major mobile networks,
GSM-MAP and TIA/EIA-41 (commonly called IS-41 or ANSI-41), for the benefit
of service providers and customers worldwide.
Following are five key technical principles that QUALCOMM believes should be
addressed by the wireless industry for successful global convergence of CDMA
proposals to achieve the highest quality, greatest spectral efficiency and most
cost effective service.
Principle #1 - A single chip rate of 3.6864 Megachips per second should
be used for the 5 MHz bandwidth
A key requirement for third generation systems is the ability to support
IMT-2000 services in spectrum allocations of 5 MHz x 2 (i.e., the network
operator has multiples of 5 MHz allocated for the transmit channel and equal 5
MHz multiples allocated for the receive channel). This is true for the D, E, and
F band Personal Communications Services (PCS) operators in the U.S. and may
also be true in other countries where regulators may partition IMT-2000
spectrum into similar-sized spectrum segments or choose to phase in the
allocation of new spectrum. The objective is to fit the wideband CDMA carrier
within this allocation, while still allowing for the requisite guard band (a narrow
bandwidth between adjacent channels which serves to reduce interference
between those adjacent channels) between this carrier and any adjacent
uncoordinated carrier.
Significant debate has ensued around the choice of the chip rate, which in
simple terms, determines the degree of spreading for the CDMA signal. The chip
rate, therefore, impacts the carrier spacing and the resultant system capacity.
Standard band CDMA TIA/EIA-95 (previously called IS-95) utilizes 1.2288
Megachips per second (Mcps) as the chip rate, resulting in a bandwidth of
slightly less than 1.25 MHz. For third generation systems, a chip rate of 3.6864
Mcps has been specified for a variety of reasons. The most obvious is the fact
that it fits well into the 5 MHz band allocation, allowing a minimum of 250 kHz
as a guard band to be provisioned on either side in order to prevent
interference with any adjacent uncoordinated operators. This is not true of the
4.096 Mcps chip rate that has been specified for the competing W-CDMA
proposal, which results in a guard band of 0 kHz, making it unacceptable from a
spectral emissions perspective. The 3.6864 Mcps value is also 3 x 1.2288, a
multiple of the current CDMA chip rate, allowing greater compatibility with the
systems now being manufactured and deployed worldwide. The cdma2000
proposal allows two approaches on the forward link, a multi-carrier option and
a direct-spread option. A chip rate of 3.6864 Mcps supports both these options
and supports a flexible third generation overlay of today's cdmaOne
(TIA/EIA-95) systems and, most importantly, equal or greater efficiency in new
green field spectrum.
The grievances voiced by the North American GSM Alliance on the third
generation convergence issue and their opposition to the 3.6864 Mcps chip
rate stem from their desire to prevent ease of transition to third generation
systems for the cdmaOne operators, and thereby avoid a competitive
disadvantage for GSM operators arising from their choice of GSM technology. It
is certainly true that use of CDMA in second generation, either initially or
following a transition, provides a competitive advantage to use of CDMA in
third generation. The power of CDMA should not be weakened in an attempt to
somewhat weaken these advantages.
The analysis for the 3.6864 Mcps based wideband CDMA system permits the
deployment of five wideband CDMA carriers in a 20 MHz x 2 band allocation,
with a minimum guard band of 250 kHz on either side of the band. The choice
of a 4.096 Mcps chip rate on the other hand allows for only four wideband
CDMA carriers in the same 20 MHz allocation, with no guard band at all. In
fact, there is significant concern that provisioning four carriers will overlap the
edges of the 20 MHz band, potentially causing interference in the adjacent
band.
In countries where cdmaOne (TIA/EIA-95) deployments exist, the operators
could mix the wideband channel with the standard band 1.25 MHz channels,
enabling them to provide a greater mix of differentiated voice and data
services. In the 15 MHz x 2 band allocation (used in Region 2 and under
consideration for new allocations in some countries), choosing 3.6864 Mcps
allows the operator to fit three wideband CDMA carriers and additional standard
band CDMA carriers. Choosing 4.096 Mcps yields the capability to support at
most three wideband carriers.
A "compromise" chip rate, 3.84 Mcps, has also been proposed. This change
from 3.6864 Mcps is driven by the advocates of competitive disadvantage, and
this change would achieve their goal without providing any technical
advantage. Thus, it is not a compromise. The 3.6864 Mcps chip rate has true
claim for acceptance in the converged standard because it builds on the 100
million or more CDMA subscribers expected by the time IMT-2000 services are
offered commercially, currently targeted for 2002.
The choice of chip rate has also been incorrectly linked by some opponents of
convergence to the issue of Intellectual Property Rights (IPR). This linkage is
erroneous and misleading since the selection of the chip rate does not make
QUALCOMM's IPR any more or less applicable. Dual chip-rate options have also
been discussed to resolve the political issue of chip rate selection. A dual chip
rate option unnecessarily complicates the implementation of the handset.
Conclusion: QUALCOMM urges the adoption of the 3.6864 Mcps as the single
chip rate for the converged standard.
Principle # 2 - Existing cdmaOne service and signaling for ANSI-41 must
be accommodated, allowing for phones both with and without SIM cards
In accordance with the International Telecommunications Union's (ITU) "Family
of Systems" concept, QUALCOMM believes that any true third generation
standard must support both existing cdmaOne services (supported through
ANSI-41) and GSM services (supported through the GSM MAP protocol) without
'interworking' (i.e., without dependence on switch-external 'hardware and
software box' implementations that inadequately adapt services of cdmaOne
systems to services of GSM and vice-versa). Dependence on interworking
removes the responsibility of accommodating existing services of either system
from the Mobile Switching Center (MSC) of the other. Current experience with
such interworking implementations has generally not been positive.
For the support of cdmaOne services, this would include the IS-127 (EVRC)
speech codec, IS-733 (13 kbps) speech codec, IS-707 data services, IS-637
short messaging, and IS-683 over-the-air service provisioning. Appropriate air
interface signaling to support ANSI-41 features and services must be efficiently
supported.
The ITU has defined a User Identity Module (UIM) function that is associated
with subscriber information and authentication in the network. The need for
UIM support is a functional requirement and is not tied to a specific physical
implementation of the function. The UIM function should be allowed to be
implemented with or without Subscriber Identity Module (SIM) cards, also
known as 'chip cards.' SIM cards have been predominantly used by GSM
systems. Elsewhere, non-card based methods for subscriber management are
widely in use, and new methods continue to be developed for future
implementation. Neither of the two physical approaches, card-based or
non-card based, should be mandated in the converged standard to the
exclusion of the other. The operator should have the flexibility to choose and
third generation systems should accommodate both.
Conclusion: QUALCOMM understands and fully accepts that GSM operators
and manufacturers require compatibility with GSM services and GSM-MAP in the
same way that cdmaOne operators and manufacturers require compatibility
with cdmaOne services and ANSI-41. QUALCOMM believes that third generation
standards should support both cdmaOne and GSM services and networks
equally. QUALCOMM also supports the evolutionary development of new
services and the convergence of cdmaOne and GSM services and networks,
and recognizes that most high burst rate traffic will be destined for the
Internet using internationally accepted Internet protocols.
Principle # 3 - Synchronous base station transmission of a shared,
time-shifted code-division continuous pilot should be used on each
forward beam
CDMA technology has successfully benefited by keeping base stations
synchronized to a common time reference. The Global Positioning Satellite
System (GPS) has been utilized for this purpose in cdmaOne deployments.
Alternative methods for synchronization are also being investigated and some
proposals have been made for use in cdma2000 systems.
The competing W-CDMA proposal stipulates asynchronous base station
operation. The rationale for proposing asynchronous systems over synchronous
systems has been stated to be 1) to avoid GPS based synchronization methods
and 2) to overcome the alleged difficulty of providing an external
synchronization source for micro base stations or pico base stations in buildings
or underground subway stations. QUALCOMM has participated in discussions on
the merits of synchronous versus asynchronous base station approaches. It is
well known that, however implemented, synchronized operation yields better
CDMA system performance with less mobile station complexity (for example, in
handoff scenarios). Further, GPS is the simplest, most economical, accurate
and omnipresent source of timing for synchronization and frequency
maintenance. Synchronization schemes other than GPS have been proposed
and can be deployed to remove any GPS related issues. However, some
companies are still insisting on asynchronous operation to the exclusion of
synchronous operation.
Conclusion: QUALCOMM believes that the third generation standard should be
based upon synchronous operation. While there are several possible
approaches for synchronous systems, the cdmaOne approach works very well
and has been extensively deployed. Further adoption will maintain compatibility
and minimize developmental risks, while achieving low cost.
In addition, QUALCOMM believes that Code-Division Multiplexed (CDM) pilots
perform better and provide greater flexibility for cell/sector wide beams, for
spot beams covering portions of cells/sectors, and for adaptive beams directed
at a single mobile station, than the Time-Division-Multiplexed (TDM) pilots
currently being proposed for W-CDMA. The W-CDMA proposal uses a dedicated
TDM pilot per mobile station, resulting in lower system capacity and exhaustion
of forward link codes utilized to support various services. The CDM approach,
therefore, yields greater capacity and flexibility.
Conclusion: QUALCOMM believes that technology modifications that yield
performance and cost benefits should be adopted wherever possible. Changing
from CDM pilots to the W-CDMA approach of TDM pilots has thus far not
demonstrated any performance or cost benefit, but to the contrary, adversely
impacts capacity and flexibility. QUALCOMM advocates full public comparative
testing to prove the purported advantages of this technology modification.
Principle # 4. - A variable rate speech codec should be used with
efficient full, 1/2, 1/4, and 1/8-rate operation with a 20 ms frame
The variable rate vocoding scheme has been a key feature of CDMA, proven to
improve system capacity and performance. This approach allows for seamless
source control, which adapts the information rate of the source and efficiently
uses different rates for different speech signal frames. The flexibility of the
variable rate vocoder is demonstrated by the ability to support full rate,
1/2-rate, 1/4-rate and 1/8-rate operation, based on the short term phonetic
character of the speech signal and on network conditions. Variable rate codecs
allow signaling information to be efficiently multiplexed onto the traffic channel
by reducing the codec's encoding rate to 1/2, thus incurring minimal voice
quality degradation for that signaling frame. The variable rate codecs can be
switched automatically to lower rate modes permitting an easy extension in
range and capacity. This approach also utilizes seamless channel control of the
encoding rate that allows the encoder to be scaled back to increase system
capacity. Variable rate encoding has been shown to be excellent for voice
storage and for packetized communication systems as well as for CDMA
allowing for seamless wireless/wireline network integration.
W-CDMA is expected to use the Adaptive Multi-Rate (AMR) codec, which uses
a degenerate and less effective method than true variable rate vocoding. The
AMR transmission rates can be adapted to the channel and source conditions
by using signaling, a far slower and cruder method of control. In contrast, the
true variable rate codec can adapt its rate on a frame-by-frame basis without
the need for signaling while incorporating the advantages of seamless channel
control. This flexibility of variable rate vocoding permits the rate to be
instantaneously lowered when the source contains greater redundancy,
increasing capacity. Furthermore, the mobile station using variable rate
vocoding can autonomously reduce its transmission rate to increase range
when at the edge of coverage (such as when inside a building).
Conclusion: Variable rate vocoding schemes have been well tested in
cdmaOne systems and can support the needs for the next generation systems.
cdmaOne systems support multiple variable rate vocoders today, permitting
flexible introduction of new improvements. QUALCOMM believes that variable
rate vocoding should be a key feature/criterion for the selection of third
generation technology. However, QUALCOMM supports the standardization of a
variety of speech codecs that utilize this beneficial method.
The cdma2000 proposal is based on a 20 millisecond (ms) frame length
parameter. QUALCOMM has significant field experience with CDMA systems and
has compared the relative performance of utilizing 20 ms frames against 10 ms
frames, as proposed in W-CDMA. Shorter frame lengths permit somewhat
shorter end-to-end delays in the system. However, in such comparisons, users
cannot perceive the reduced delays, which are still on the order of 50-70ms.
On the other hand, frame lengths of 20 ms are clearly more efficient in terms of
overhead, and support higher system capacity, compared to the use of the
smaller frame lengths.
Principle # 5. The reverse link waveform should have low energy in the
audio band due to amplitude modulation and should support
enhanced-range low-rate phones.
QUALCOMM believes interference with hearing aids and other devices must be
minimized. To minimize interference, the radio frequency (RF) waveform must
have low- energy in the modulation components, which are in the audio band.
In general, CDMA systems permit the design of air interfaces that minimize the
interference into the audio band.
Conclusion: Although QUALCOMM believes that this issue has been addressed
in the W-CDMA proposal, the Company stresses the importance of this
consideration as future modifications are discussed.
Summary:
This white paper is intended to clarify QUALCOMM's position on the five
technical principles that were communicated to some of the world's key
standards bodies. As part of the process, QUALCOMM also looks forward and
expects to converge the W-CDMA and cdma2000 proposals in other areas to
achieve a high quality, cost effective third generation system to meet the
needs of global wireless users in the 21st Century. QUALCOMM recognizes that
many companies are actively working on these standards and that many
valuable innovations are being introduced and should in many cases be included
in the standard, after proper study and test. Intentional introduction of a
competitive disadvantage must not be one of the criteria for acceptance.
QUALCOMM has an extensive CDMA patent portfolio with over 130 CDMA
patents issued and approximately 400 patent applications pending in the United
States, Europe, Japan, Korea, China and elsewhere around the world. Prior to
the adoption of the cdmaOne standard, QUALCOMM committed to license its
essential patents for such standard on reasonable terms and conditions free
from unfair discrimination. QUALCOMM has made good on its commitment as
evidenced by the fact that QUALCOMM has licensed its essential patents for
use in cdmaOne applications to more than 55 companies including nearly every
major telecommunications manufacturer in North America, Europe and Asia.
QUALCOMM believes it has essential Intellectual Property Rights (IPR) for
W-CDMA, and it intends to license these patents on reasonable terms and
conditions free from unfair discrimination for a single converged IMT-2000
standard, or, if not achieved, only for cdma2000. In addition, QUALCOMM has
extended or is preparing to extend these bilateral agreements, and the
Company intends to review its royalty rates within the context of the market
size that will be achieved by a single converged standard.
QUALCOMM has no intention of generally licensing its essential patent portfolio
for any IMT-2000 standard (such as W-CDMA) that is purposefully made
incompatible with cdmaOne and ANSI-41 without providing a material benefit to
the industry.
Headquartered in San Diego, QUALCOMM (NASDAQ: QCOM) develops,
manufactures, markets, licenses and operates advanced communications
systems and products based on its proprietary digital wireless technologies.
The Company's primary product areas are the OmniTRACSr system (a
geostationary satellite-based, mobile communications system providing
two-way data and position reporting services), CDMA wireless communications
systems and products and, in conjunction with others, the development of the
GlobalstarT low-earth-orbit (LEO) satellite communications system. Other
Company products include the Eudora Pror electronic mail software, ASIC
products and communication equipment and systems for government and
commercial customers worldwide. For more information on QUALCOMM products
and technologies, please visit the Company's web site at
<http://www.qualcomm.com>.
QUALCOMM, OmniTRACS and Eudora Pro are registered trademarks of
QUALCOMM Incorporated. Globalstar is a trademark of Loral QUALCOMM
Satellite Services, Incorporated. cdmaOne and cdma2000 are trademarks of
the CDMA Development Group.
CDMA Digital Technology | Corporate Overview | Investor Relations | 1998 Press Releases || QUALCOMM Home |
