Clarity & Understanding: The High-speed WLAN standards debate
March 12, 2002
This is a detailed analysis of the technological and economic obstacles that the .11g draft spec faces, particularly in relation to .11a production and implementation.
Recent approval of the 802.11g wireless local area network (WLAN) draft specification has led to some confusion in the industry regarding the status of high-speed WLANs. The draft specification was presented in November 2001 by the 802.11g Task Group (TGg), which has been developing an addendum to the existing 802.11b 2.4-GHz WLAN standard for a high data-rate and backward-compatible extension.
Since the IEEE approval of draft 802.11g specification, published articles in industry trade publications as well as communications originating from industry participants have contributed to the confusion by incorrectly stating the status of the proposed 802.11g standard, embellishing what the draft specification offers to the WLAN market and casting doubt on the future market prospects of the 802.11a standard. Currently, there are two approved IEEE WLAN standards- 802.11b (operating in the 2.4-GHz ISM band) and the higher speed 802.11a (operating in the 5-GHz band). The draft 802.11g specification is an extension of the 802.11b standard that brings Orthogonal-Frequency-Division Multiplexing (OFDM, the modulation scheme used by 802.11a) to the 2.4-GHz ISM band.
In an effort to distill the issues and realistically depict the industry's drive to high-speed WLAN, this paper provides information that should be considered for an informed position in the current high-speed WLAN standards debate.
For the sake of clarity, it is vital to understand the status of the 802.11g draft specification in the context of the IEEE's process to ratify new standards. In order for 802.11g to be successfully ratified, there are numerous engineering issues that must be addressed before gaining the approval of organizations such as the IEEE, FCC, and WECA. As a means of comparison, it is useful to examine the history and current status of 802.11a standards-based products, the standard most recently ratified by IEEE. As a result of the performance, market support, and time in the market, the 5-GHz 802.11a standard may ultimately have an advantage over new 802.11g products.
2. Status of 802.11g Ratification
At the January 2002 IEEE Standards meeting, the 802.11g Task Group (TGg) received approval for a draft 802.11g specification, thereby completing the second step of the standard ratification process. (See "Understanding the IEEE Standards Ratification Process," at the bottom of this document for a description of the 5-step process to ratify a new IEEE standard or addendum.) TGg expects to complete the remaining 3 steps and achieve ratification by Q1 2003.
3. 802.11g: The Road to Volume Production
Since it is still a draft specification, it will be some time before Wireless Ethernet Compatibility Alliance (WECA)-certified 802.11g products are delivered to the market in high volume. In fact, a number of critical issues must be addressed before achieving IEEE ratification and FCC regulatory approval of the 802.11g standard: IEEE ratification and FCC approval, followed by WECA-recognition and certification of 802.11g products, and the start of true volume production.
3.1 Issues For Ratification
Moving the 802.11g specification from its current draft format to a ratified standard will require a significant amount of engineering effort by the TGg to identify and implement solutions for a number of already identified technical issues, including the absence of a full solution for backward compatibility in operation with legacy 802.11b, co-existence in the 2.4 GHz ISM Band, and potential conflicts in the media access controller (MAC).
3.1.1 Backward Compatibility
The draft 802.11g specification does not define the technical solution required for backward compatible operation with existing 802.11b infrastructure. There are a number of areas in which the draft specification should be improved.
To be backwards compatible with 802.11b, the 802.11g client must be capable of receiving and detecting the short preamble of the 802.11b header at a rate of 1Mbps. Unfortunately, legacy 802.11b products in the market do not have the capability of detecting and recognizing the comparable 802.11g packet. Currently, the 802.11g draft specifies that the Request to Send/Clear to Send (RTS/CTS) "listen before talk" mechanism be used to combat this problem and reduce the number of collisions during transmission. However, if the timing between the RTS/CTS frames is not precisely controlled, then the system may suffer in overall system throughput, especially at the higher data rate of 54Mbps. To successfully support interoperability, the specification needs to address this technical limitation.
In addition, for reliable interoperability, the specification must also define the protocols required for negotiating between modulation schemes and the different modes of operation associated with various user models. For instance, the parameters for switching between Complementary Code Keying (CCK) and OFDM-such as speed of switching (per packet basis or per session basis) as well as the modulation discovery process-still must be defined in any ratified 802.11g specification.
In addition to the technical issues, there are practical infrastructure issues that must be considered by consumers managing existing 802.11b networks who may be considering upgrading to 802.11g in the future. It has been previously communicated that legacy 802.11b infrastructure can be seamlessly upgraded to support the future 802.11g standard. The reality is that the radio performance requirements of draft 802.11g specification, already far exceed that of current generation 802.11b radios, making existing 802.11b radios obsolete. As a result, moving to 802.11g will require field upgrades of existing 802.11b infrastructure.
Having positioned 802.11g as a solution for backward compatible operation with 802.11b, TGg will be held to task by the standards community to deliver on this promise. Bringing the current draft 802.11g specification to one that addresses solutions for true backward compatibility will require considerable engineering effort and debate among standards bodies, easily delaying ratification of the 802.11g standard beyond the original Q2 2002 target to Q1 2003 or later.
3.1.2 Co-existence in 2.4-GHz ISM Band: "Friendly" or "Un-Friendly" 802.11g?
The 2.4-GHz Industrial-Scientific-Medical (ISM) band currently hosts Bluetooth and 802.11b along with numerous other devices, such as cordless telephones and microwave ovens. It is becoming an increasingly crowded spectrum, resulting in unreliable operation of some products already operating in this band.
The impact of adding the OFDM modulation scheme (as proposed in 802.11g) to an already crowded spectrum requires further investigation. There have been no definitive studies conducted to determine how legacy 802.11b client and access point equipment will handle the energy produced by the new high data rate modes of 802.11g. At a minimum, there is likely to be significant degradation in the performance of legacy 802.11b networks.
This issue of co-existence of OFDM in the crowded 2.4-GHz band must be taken into consideration by the TGg to ensure that a "friendly" specification is ultimately developed. In the event an "unfriendly" 802.11g specification is passed, it is likely that 802.11g products will have a negative impact on the performance of other wireless products operating in the 2.4-GHz ISM band.
3.3.3 Impact on 802.11 Media Access Controller
One of the most intriguing unknowns is what impact 802.11g will have on the 802.11 Media Access Controller (MAC) and, specifically, on the MAC extensions currently being developed for improving Quality of Service (QoS; 802.11e). To avoid difficulties in this area, TGg must determine solutions to QoS problems that 802.11g client applications may encounter while roaming.
In addition, the group needs to determine a solution to the apparent violation of QoS protocols by dual-mode 802.11g access points. For instance, the work coming out of 802.11e for QoS assumes a fixed delay in the data processing path. This is violated by multi-modulation modes due to their inherent processing delays. Also, the need to back off transmissions via RTS/CTS will increase the jitter in data streams. The engineering effort required to address this issue is likely to push ratification of the 802.11g standard to Q1 2003 or later.
3.2 Issues for FCC Regulatory Approval
The 802.11g draft specification proposes moving the 5-GHz 802.11a signal down to the 2.4 GHz band (downbanding). This will cause out-of-band emissions that will violate the 2.4 to 2.4835 GHz emissions limits specified by the Federal Communications Commission (FCC). If this issue is not resolved, this would prevent 802.11g products from achieving FCC regulatory approval for operation in the 2.4-GHz ISM band.
To bring the 802.11g spectral mask within regulatory allowances, a number of revisions to the draft 802.11g standard are possible. For instance, the output power of the 2.4-GHz OFDM signal could be lowered (from +20 dBm to +10 dBm), which would reduce range. Another possible solution could involve altering channel spacing, which would cause more interference with the operation of legacy 802.11b networks and reduce the number of channels, thus reducing the already limited channel capacity of future 802.11g networks. (see "802.11g Spectral Mask Violations, Potential Solutions, and Trade Offs" at the bottom of this document .)
3.3 Issues for WECA Recognition and Certification
A major factor in the recent market success of 802.11b WLAN products was the industry's efforts to develop device interoperability testing and marketing programs. The Wi-Fi brand and certification program for 802.11b products, developed and managed by the WECA, was instrumental in giving consumers the assurance they needed that WLAN devices would be fully interoperable. It is fair to expect that WECA will use the same approach to recognize and certify other WLAN products. In fact, it has already begun this process for 802.11a.
The first step in achieving WECA certification of 802.11g products would be the development of the certification program itself. This will require discussions between WECA and 802.11g proponents to develop the necessary administrative, marketing, and testing (compliancy to a PICS Proforma defined in the ratified 802.11g standard) programs. An example of such as program is the Wi-Fi5 certification branding for 802.11a 5-GHz WLAN products that was announced by WECA in the fall of 2001.
With a program in place, certification of 802.11g products will begin when 802.11g-compliant semiconductor solutions are available from at least two different semiconductor providers. Assuming the 802.11g standard is ratified by Q1 of 2003, it is reasonable to expect that WECA certification of 802.11g products will begin in Q3 2003 with volume production of WECA certified products beginning Q1 2004.
3.4 Achieving Volume Production
After IEEE ratification, FCC approval, and WECA certification, the final hurdle for achieving volume production of 802.11g products is ensuring that these products deliver a satisfying end-user experience. This will drive mass-market demand. To be successful, WLAN system developers must identify and take advantage of WLAN integrated circuit (IC) solutions that consume low DC power, support a low total system cost, and offer superior performance.
3.4.1 Low Power Consumption
The draft 802.11g specification proposes the use of 802.11a OFDM in the 2.4-GHz ISM band. As is the case at 5-GHz, traditional 2.4-GHz OFDM modems consume significant power due to the high peak-to-average power ratio (PAPR) of OFDM radio signals. High PAPR results in very inefficient power amplification, increasing power consumption and heat dissipation.
The high power consumption of traditional architectures combined with the overhead of two baseband processors may produce a power profile for 802.11g that does not fit within the limits of PC-Card client cards, thus requiring the RF output power to be scaled back. The high power consumption of traditional OFDM modems will also have a negative effect on the mobile 802.11g end-user due to reduced battery life and higher device casing temperature, resulting from greater heat dissipation. In addition, this increased heat dissipation may place packaging constraints on the design of 802.11g, which could increase the size and cost of 802.11g products.
In order to provide an optimal end-user experience and drive market demand, designers of 802.11g systems must identify a semiconductor system solution with an innovative OFDM modem architecture that decreases the PAPR, thereby reducing power consumption. IceFyre Semiconductor, for instance, is currently developing an OFDM modem design that can be applied to address the demand for low power in future 802.11g WLAN products.
3.4.2 Low Total System Cost
Presently, the total system cost of an 802.11g solution is unknown. However, it is possible that implementing an 802.11g traditional semiconductor system solution that features interoperable CCK and OFDM modulation and delivers acceptable performance may not be cost effective.
Delivering a cost-effective 802.11g semiconductor system solution is vital to ensuring that 802.11g end products offer the price/performance ratio necessary to drive consumer demand over mature 802.11b products and proven 802.11a products. Therefore, it is crucial to the success of 802.11g products that semiconductor system designers explore new technologies and architectures in order to minimize cost early in the product life cycle.
3.4.3 Superior Performance
The current draft 802.11g specification provides for raw data rates of 54 Mbps and a 100-m range-numbers that are comparable to the 802.11a standard. In terms of actual data throughput, single-mode 802.11g networks will handle approximately 27 to 30 Mbps, which is comparable to the performance of single-mode 802.11a and dual-band 802.11a+b networks. However, for dual-mode 802.11g networks in which 802.11g and 802.11b equipment coexist, the throughput of the 802.11g components drops to 20 Mbps or less due to the channel time sharing required to support dual-mode operation.
In order to coexist with legacy 802.11b systems, 802.11g clients must delay their transmissions to accommodate 802.11b clients, which cannot "see" them. At a given data rate, the effective throughput is proportional to the on-air time. The channel time-sharing operation of dual-mode 802.11g reduces effective throughput, and as the number of 802.11b clients increases, the air time available for 802.11g clients decreases, which leads to a larger drop in throughput. It is believed that in the presence of 802.11b clients, the throughput of dual-mode 802.11g networks will not exceed that of a single 802.11b network.Figure 1 shows a comparison of throughput for a range of WLAN networks including 802.11b only, 802.11g only, dual-band 802.11a+b and dual-mode 802.11g.
Figure 1: Throughput Comparison of WLAN Networks
Unfortunately, with a maximum data throughput of 20Mbps or lower, dual-mode 802.11g offers poor support for audio/video applications such as high definition television (HDTV). As a result, consumers or network administrators interested in deploying an 802.11g-based high speed WLAN in home, enterprise, or public access environment to support high bandwidth applications would be required to upgrade all Access Points and Network Interface Cards (NICs) to 802.11g in order to maximize network performance.
4. 802.11a: First Mover Advantage
4.1 IEEE Ratification and WECA Certification
The IEEE 802.11a standard for 5-GHz WLANs was ratified in 1999. It achieved WECA recognition in the fall of 2001 with the finalization of the Wi-Fi5 brand. A draft of the certification program for Wi-Fi5 will be presented in March 2002, and the program is expected to be finalized by Q3 2002. WECA anticipates certification of 802.11a products to begin in Q4 2002 with volume shipments occurring in Q1 2003-at least one year ahead of similar 802.11g products.
4.2 Volume Production
Non-WECA certified 802.11a-compliant products began shipping in limited production runs in Q4 2001. Although current or announced semiconductor system solutions will serve as effective platforms for these limited production-run products, innovative semiconductor systems are required to deliver low power consumption, low total system cost, true 802.11a performance, and backward compatible operation with 802.11b, all of which are essential to deliver the end-user experience that will drive volume production of 5-GHz WLAN products.
Fortunately, IceFyre Semiconductor is meeting market demands for a low power 802.11a semiconductor system with its innovative 5-GHz OFDM Modem that significantly reduces power consumption while cost-effectively delivering performance that meets or exceeds 802.11a specifications.
4.2.1 Low Power Consumption
Similar to concerns in the proposed 802.11g specification, the OFDM modulation technique used in 802.11a traditionally consumes significant power due to a high PAPR. The high power consumption of traditional architectures may often produce a power profile for 802.11a that does not fit within the design limits of PC-Card client cards, thus requiring RF power output to be scaled back. An innovative technology and architecture, such as the one employed by IceFyre Semiconductor, is required to successfully lower power consumption.
4.2.2 Low Total System Cost
To be competitive with current 802.11b products, 802.11a semiconductor system solutions must strive for low total system cost. Current 802.11a semiconductor system offerings based on traditional OFDM modem designs are typically multi-chip solutions that feature high external component counts, resulting in a large Bill of Materials (BoM).
Several companies competing in 802.11a chipset market are working on lowering the system BoM. Some are pursuing a risky Direct Conversion Zero-IF architecture to reduce the number of external components. IceFyre's innovative computational-based OFDM architecture includes a receive architecture which achieves all the touted benefits of a Zero-IF design, but avoids the difficulties in a practical realization of Zero-IF. It also reduces the number of external components by 25% over competing designs. All of this contributes to a lower system cost.
4.2.3 System Performance
To gain a competitive advantage over current 802.11b products, 802.11a products must deliver true high-speed WLAN performance including 54 Mbps and range of 100m. 802.11a WLAN technology used in either single-mode 802.11a or dual-band 802.11a+b networks offers data throughput of about 30 Mbps, which is sufficient for use in A/V applications in the home, enterprise, or public access environments.
With 8 channels for data networking, the 802.11a standard provides the network capacity that WLAN administrators in the Enterprise and Public Access environments will need to cost-effectively deliver the quality end-user experience that will encourage mass adoption of high-speed WLAN devices.
Experience using current 802.11b products in real world conditions, such as crowded boardrooms, has shown the industry that designers should strive to exceed 802.11a performance specifications in key areas such as Receive (Rx) Sensitivity and Error Vector Magnitude (EVM). IceFyre took these issues into account and developed an innovative OFDM modem design that provides true 802.11a performance while still delivering low power consumption and low total system cost.
4.2.4 Backward Compatibility
Backward compatibility of 802.11a with 802.11b is a key design requirement for next-generation high-speed WLAN technologies. To address this market requirement, numerous semiconductor companies are working on dual-mode 802.11a+b solutions. It is anticipated that WECA-certified dual-mode 802.11a+b products will begin shipping in volume in early 2003, a year or so ahead of comparable of 802.11g products. Based on this availability, dual-band 802.11a+b solutions are well positioned to lead the drive for high-speed WLAN deployments.
4.4.3 Global Market Acceptance of 5-GHz WLANs
Industry participants have noted that the total market size available to 802.11a is currently limited due to the fact that it is not a globally recognized standard. To address this issue, WECA and ETSI/ERC are actively lobbying to have the 5.15 - 5.35-GHz band dedicated for WLAN use in Europe. These efforts are having an impact, mostly because of the weak market acceptance of HiperLAN2 (HL2).
In terms of the Asian market, 802.11a has already been adopted in Japan-but with only 4 channels of bandwidth (as opposed to 8 in North America). Current efforts are underway to license the 4.9-GHz band in Japan and thereby increase the number of channels for 802.11a. In addition, Japan has also adopted a standard similar to HL2 called HiSWANa, but, with the weakening of the HL2 standard, it is believed that Japan will ultimately favor the 802.11a standard.
In addition, at the World Radio Congress (WRC) in 2003 (WRC-O3) it is widely anticipated that the 5.15 to 5.35 GHz U-NII band will receive global endorsement for dedicated use by WLAN applications (in addition to extra band in 5.45 to 5.75 GHz).
Some organizations have commented that spectrum priority issues may create problems for the operation of 5-GHz WLAN technology, since radar and satellite applications currently have "primary use" designation in the 5GHz bands used by 802.11a. This means that 802.11a equipment with "secondary use" designation must mitigate interference to radar and satellites in these bands. The use of Dynamic Frequency Selection (DFS) and Transmit Power Control (TPC) is under review by spectrum regulatory boards and it is expected that the WRC-03 will declare global acceptance of these measures, thus ensuring interference-free operation of 802.11a WLANs. In support of these activities, the IEEE 802.11h Task Group expects to ratify the DFS and TPC extensions required for operation in Europe by Q3 2002.
Achieving global endorsement of the 5.15 to 5.35 GHz band for dedicated use by WLAN applications will significantly increase the market size and resulting volume production requirements of 802.11a client and access point products. These diverse efforts are working towards the goal of making 802.11a a globally recognized WLAN standard in 2003.
4.4 Market Advantages
802.11a products will likely be on the market for at least a year before 802.11g 2.4-GHz products become available in volume. When this is combined with their performance and global market momentum, 5-GHz WLAN products are expected to have a solid advantage in all markets.
The 802.11a standard will enjoy a strong short and long-term position in the enterprise market due to its strengths in the following key areas:
- Data throughput advantage of dual-band 802.11a+b over dual-mode 802.11g;
- 12 month time-to-market advantage over 802.11g;
- 802.11a's 8 channels provides better support of dense user enterprise environments than 802.11g, which is limited to a maximum of 3 channels and possibly fewer to solve Spectral Mask violations;
- No co-existence problems with applications in the 2.4-GHz ISM Band will allow network managers to offer a full compliment of wireless productivity tools.
The 802.11a standard will enjoy a strong short and long-term position in the residential market due to its strengths in the following key areas:
- Data throughput of dual-band 802.11a+b meets the requirements of A/V applications stronger than dual-mode 802.11g;
- No co-existence problems with applications in 2.4-GHz ISM Band.
Public Access Market
The 802.11a standard will hold a strong position in the Public Access market due to its strengths in the following key areas:
- Data throughput advantage in dual-band 802.11a+b over dual mode 802.11g;
- 12 month time to market advantage over 802.11g;
- 802.11a's 8 channels provides better support of dense user enterprise environments than 802.11g, which is limited to a maximum of 3 channels and possibly fewer to solve Spectral Mask violations;
- No co-existence problems with applications in 2.4-GHz ISM Band.
In order to provide 802.11g-based products with the opportunity to achieve success in markets in which 802.11a-based products have performance and operational advantages, the 802.11 TGg must work quickly to resolve the technical, functional, and regulatory issues facing the current draft specification. If ratification of the 802.11g standard is delayed significantly beyond Q1 2003, or if the ratified standard adversely affects other 2.4 GHz ISM Band applications or fails to provide the performance demanded by users (throughput in dual-mode operation, range, network capacity), 802.11g may find it difficult to gain traction in the market.
|Understanding the IEEE 802 Standards Ratification Process
The IEEE has established a rigorous process through which standards are debated, reviewed, revised, and approved. There are six major steps that form the standard development and approval process as described below (see Figure 2). (For more information, visit www.IEEE.org.)
Step 1: Establish a Working Group (WG) or Task Group (TG) under a working group authorized to develop a standard (or an extension to a standard) as defined by a Project Authorization Request (PAR). The PAR defines the market and technical requirements necessary to meet the demands of the industry. The 802.11 Task Group g (TGg) and the 2.4-GHz High Data rate PAR were established in September 2000.
Step 2: The TG or WG creates draft text for a standard or an addendum to an existing standard. This is conducted by soliciting and selecting the best technical proposal followed by the generation of supporting text so that equipment manufacturers, silicon suppliers, and software developers will be able to develop product. Then, the TG or WG seeks approval of the draft specification. TGg received approval of its draft specification in January 2002.
Step 3: TG or WG undertakes WG Letter Ballot review of the draft standard. Members are able to submit editorial and technical comments, questions, or concerns about the draft specification for consideration and possible remedial action by the TG or WG. During this period, the draft specification may be circulated many times as it is revised by the TGg. This letter ballot review lasts 40 days. TGg expects to complete this first letter ballot by April 2002.
Step 4: A sponsor Letter Ballot is conducted by, The Computer Society, the sponsor of the IEEE 802 working group. This broadens the technical review pool beyond the IEEE 802 working group. The draft specification may be circulated a number of times as it is revised to incorporate changes requested by the Computer Society and agreed to by the WG or TG. To move the draft text to final stages of approval by the Executive committee, the approval must reach greater than 75%. Sponsor Letter Ballot is expected to last 40 days, and the TGg expects to complete this by July 2002.
Step 5: Submit to the IEEE 802 Executive Committee (Ex-Com) for final review and approval, and then to the standards board for printing. TGg expects to submit the 802.11g standard in Q1 2003.
| 802.11g Spectral Mask Violations, Potential Solutions, and Trade Offs
The current 802.11a spectral mask requires out-of-channel energy to be at least 40 dB below the on-channel signal, as opposed to the 802.11b spectral mask requirement of 50 dB (see Figure 3). This means that OFDM, with the same power as CCK, will violate FCC out of band emissions in channels 1 and 13 in 802.11b or 802.11g operation.
Fig. 3 Spectral Mask of 802.11a OFDM signal downbanded to 2.4 GHZ ISM Band