'Disaster.' 'Miserable failure.' These are just a couple of the terms Garth Freeman, CEO of Buzz Broadband, used to describe his experience with Wimax at a conference in Bangkok last March.
Freeman warned the audience that Wimax 'may not work' and that it was still 'mired in opportunistic hype'. The source for Freeman's frustrations was Buzz Broadband's own experience with its fixed Wimax network. Freeman said that the network's non-line-of-sight performance was 'non-existent' beyond just 2 km from the base station and that the indoor performance decayed at just 400 meters. He also said that the latency rates reached as high as 1000 milliseconds - values that make services like VoIP virtually useless.
Airspan, Buzz Broadband's equipment provider, was quick to respond. CMO Declan Byrne released a statement saying that Buzz Broadband had deployed Airspan equipment that is also installed in over 100 other fixed Wimax networks, but that Buzz had traded off range for cost by using less-expensive microcell base stations. As for VoIP QoS, wrote Byrne, '[AirSpan's] MicroMAX certainly offers appropriate QoS for wire-line quality voice support, but, as an access technology, can only do so for the portion of the link between the user device and the base station. In the case of Buzz Broadband, we know that there were significant under-provisioning issues in the core network which connected the Airspan equipment to the Internet.'
Sprint was quick to separate itself from Buzz Broadband, pointing out that while Buzz Broadband was operating a fixed Wimax (802.16d) service at 3.5-GHz - which requires line of sight - Sprint was using mobile Wimax (802.16e) in the 2.5-GHz band, 'which does not require LOS and has better building penetration.'
Still, the Buzz Broadband flap raises a legitimate challenge that Wimax players face in offering VoIP services. And by many accounts, it's becoming increasingly imperative that they do offer VoIP. At every major and minor Wimax conference in the last year, operator licensees, vendors and analysts alike have maintained that, in many markets, a voice service is essential to Wimax's success. In July, ROA Group analyst Ku Kang said in a report that one reason KT's WiBro service in Korea has seen sluggish take-up to date is that it lacks a killer app, and that app may well be VoIP, which WiBro (and Wimax) needs if it wants to compete against HSDPA and LTE.
Whatever one makes of Buzz Broadband's technology decisions, the fact of the matter is that VoIP over Wimax is a real challenge, says Tom Flak, senior operations VP and chief marketing officer for SOMA Networks - but it's a mistake to focus just on the air interface.
'Getting voice to work over the interface is the tricky part of the problem but it's not the whole problem,' Flak says. 'The technical challenges are very much the same in doing voice over Wimax as they have been for the last ten years in doing voice over broadband wireless in general. Wimax does provide some built-in quality-of-service channels to carry voice. But voice quality requires an end-to-end solution. In general delivering an access system on an end-to-end basis is still a technically challenging task.'
Some of the challenges are relatively easy to overcome. Sprint's Polivka alluded to one when he stated that Buzz Broadband's network was operating at 3.5 GHz, while Sprint is operating at 2.5 GHz. A network can not defy the laws of physics: signal loss is not only a function of the distance between the handset and the base station, but also a function of the frequency of the system - actually the square of the frequency.
So Wimax systems that use a 2.5-GHz radio system instead of a 3.5-GHz system are not just reducing their chance of signal loss in direct proportion to the difference in frequencies, but by a much larger value.
However, there is one challenge - albeit not a fatal one - that is proving to be difficult to resolve.
Power-hungry handsets
One of the key selling features used to move a customer to Wimax is offering VoIP service in addition to broadband data services. This certainly was the case with Buzz Broadband.
Of course, customers are naturally assuming that since VoIP service is part of the Wimax service, then Wimax handsets will soon be available. However, handset designers are beginning to worry whether they will be able to design a handset with long-enough battery life. The hang-up is the power amplifier used in the handset's transmitter. Put simply, it's not very efficient.
While it's not a big deal when using a PC or a USB Wimax-phone tethered to a PC's massive battery, it does becomes a big deal in handsets. The power consumed by the power amp is currently about 1W to 1.5W, which cuts significantly into the life of a handset's battery.
One measure of a power amp's efficiency is its power added efficiency (PAE) rating. PAE is the ratio of the power going into the power amp to the output power of the transmitted signal. A high PAE means that most of the input power from the battery is converted to signal output power. A low PAE means that a large portion of the input power is lost to heat or otherwise wasted.
The PAE for a GSM phone is 50% to 55%. For a WCDMA phone, it's 40% to 45%. For a mobile Wimax phone, it's 10% to 20%.
The culprit for the low PAE is the transmission method used in Wimax: orthogonal frequency division multiplexing. OFDM uses multiple carriers. That imposes a major load on the power amp, because OFDM requires that the power amp deliver very high linearity over a wide dynamic range. That can only happen by boosting power consumption.
The dilemma that designers find themselves in is that Wimax requires a signal power output level of about 250 mW. Yet, for the power amp to operate in its linear operating range so it can handle the multiple signals being multiplexed at a given time, its actual power output needs to be about 2.5W. That is, the power amp consumes 2.5W of power from the battery to produce a 250 mW signal.
Power amp manufacturers are scrambling to boost the PAE of their devices. Their immediate goal is to boost PAEs to 20-25%, and then to 40-45% levels. However, handset manufacturers can't wait - the launch date for mobile Wimax is fast approaching.
Power/data tradeoff
So designers have two choices in using today's power amps. One: lower the data rate for the uplink. Two: lower the output power of the handset.
By using simpler modulation technologies - QPSK instead of 64-level QAM for example - and fewer carriers, then the designers can reduce the linearity requirements of the power amp.
Lower linearity requirements means that higher efficiency (lower power consuming) amps can be used. However, the data rate will have to be cut to between 5% and 20% of the downlink data rate for this option to provide tangible power savings.
The second choice, that of lowering the output power of the handset means that the needed power from the battery is lowered. That's the good news. The bad news is that with lower signal power, then the handset needs to be closer to the base station for a good connection. In other words, the effective radius of the base station is reduced, necessitating more base stations in a given geography, which increase the operator's equipment costs.
Whether both options are acceptable to operators and customers for the next two or three years, during which time power amp manufacturers push to develop high-efficiency power amps, remains to be seen.
By the way, the power-amp problem is not unique to mobile Wimax. The OFDM-based Long-Term Evolution (LTE) standard was also faced with the same problem. However, 3GPP 'solved' the problem by deciding not to use OFDM on the uplink transmissions.
On the other hand, the IMT-Advanced 4G mobile communications standard is almost sure to use OFDM. If the power-amp problem is not solved in Wimax, then it will come back to trouble designers again in the next generation standard.