Your browser is not supported. Please update it.

17 November 2011

Quantenna ups the ante with first move in next gen 802.11ac WiFi wars

The world’s major WiFi chip makers go green at the mention of the name Quantenna. Not only has it already upset the established WiFi order of things by adopting the only full implementation of 802.11n WiFi, complete with 4 x 4 MIMO and offering a system truly adept at delivering HD video around a home, but now it has adapted its technology to get the first 802.11ac chip out, striking the first PR blow in a battle for hearts and minds in home gateways over the next 4 to 5 years.

All the analyst firms who focus on this area have categorically stated that there are no reasons to arrive late at what is expected to be the next level of bulk WiFi chips ‘ the standard will not be absolutely final until the end of 2012, but all the hard decisions have been taken and pre-standard products were always going to lead the standard by 12 months. Both ABI and In-Stat picked out a host of potentially first mover chips including Broadcom, Intel, Qualcomm Atheros and Texas Instruments ‘ but none said that Quantenna would lead the pack.

As we move towards a world where 1 billion WiFi chip ship each year, companies like Broadcom and Atheros have perhaps been prepared to cede small chunks of specialist WiFi ground to specialists like Quantenna, opting instead to soak up the low end volume in 802.11n, and placing the odd bet, as Atheros has done, on WiGig/802.11ad chips in a partnership with Wilocity.

WiFi technology is becoming increasingly important. Qualcomm, despite already having a decent understanding of WiFi and building its own low power version for mobile chips, acquired Atheros early this year for $3.1 billion, saying it would move into neighboring territories such as home networks. In direct response to this in March Taiwan’s MediaTek, already chasing down Qualcomm in the mobile baseband market, decided to acquire Ralink in a copycat move. Ralink already supplies similar chips to Atheros for home Wi-Fi equipment, as well as DSL and USB components, and has been involved in many efforts to extend the 802.11 standard’s remit, including the WiGig group for 60GHz connections. At the time the MediaTek promised that it would be one of the first to arrive at next generation technologies.

The precise importance of the 802.11ac market is in some question however, with one analyst firm saying it will be a mere 350 million device market by 2015, still lagging behind 802.11n, which will make up the bulk of the remaining chips, while others say that it will overtake 11n in short order, dominating by 2014.

The reason for this confusion and for the rapid Quantenna move is that slowly all general purpose WiFi chips will move to being hybrids able to support both 11n and 11ac. The difference is being able to use 40 and 80 MHz channels instead of just 20 MHz, which is mostly a radio issue, and both will require an increasing number of antennas and radios, to utilize multiple channels, and are likely to end up at least 4 x 4 MIMO. However 11n receiver chips will get no advantage at all out of working with 11ac transceivers chips, because 11n can only work with 20 MHz channels, and so won’t improve their throughput. This means an entire generational change will have to subsequently happen at the client device end of the market, and any device which wants high throughput, as high as the 2Gbps theoretical PHY maximum that 11ac can give, will need to have an 11ac chip.

Dr. Sam Heidari, CEO at Quantenna told us, ‘AC has a lot of different options and we have taken up some of those major options, to use the wider 80 MHz band and to use 256 QAM encoding. We will also use multi-user MIMO for the first time. Our existing 11n chip designs are already backward compatible with these changes.’ The ability to use MIMO to two different receivers at a time is managed by each receiver filtering out data which is not intended for it, Heidari said.

‘We will be giving samples now to a couple of select customers and in early 2012 we will be making these chips generally available,’ Heidari added . In the real world Heidari thinks that a multiple of 2 to 2.2 times the performance of its own 11n chips is likely with 11ac. ‘We have aimed our 11n product at TV operators, and this 11ac product is aimed more at retail and consumer electronics companies, and at first it will mostly go into home gateways. Gradually CE client devices will adopt 11ac, but a lot of operators don’t want to use that extra bandwidth for their TV products,’ he said, implying that existing Quantenna chips would continue selling for some time.

Quantenna says that TV operators have been able to achieve 100 Mbps in real world performance throughout a home using its 11n chips, in 95% of all trial homes, regardless of the types of walls or architecture in a building. ‘We don’t see any need to partner with fixed line technologies in order to reach all of a home, such as G.hn or HomePlug. Our WiFi is good enough on its own and slowly operators will come to accept that. About 5% of homes might use such technologies side by side, but the rest can simply move to WiFi.’ Applying his real world multiplier for 11ac chips, that means that most homes with 11ac routers with 4 x 4 MIMO might just as well achieve 200 Mbps to 220 Mbps in real world conditions.

Recently Quantenna announced that Telefonica had taken a strategic investment in the company and confirmed to us that this was an investment door left open on its last round of funding, specifically for a trade investor, rather than an entirely new round. ‘The Telefonica deal will lead to a lot of new relationships with equipment vendors,’ he added.

Given that Telefonica supports millions of home gateways around the world, especially in Europe and Latin America, that deal may stimulate more 11ac chip sales than 11n.

Quantenna chooses to accept the version of events from In-Stat which has it that after small shipment volumes in 2012 that IEEE 802.11ac will emerge as the dominant Wi-Fi protocol by 2014. And here it points to a critical lead it has in the on-coming war, and perhaps a reason why it is first to market. It says that only a niche subset of 802.11ac will be single-band, just the 5 GHz band, most of them will be 802.11n/802.11ac dual-band chipsets and many will need MIMO.

The more established Wi-Fi chipset vendors such as Broadcom and Qualcomm Atheros will have to catch up the MIMO gap that Quantenna has created ‘ both have 3 x 3 MIMO but their volume products are at 1 x 1 and at 2 x 2 ‘ or come out with WiFi devices which perform at a considerably lower spec. We will see in the coming months which set of predictions is true ‘ either these vendors will come to market in the next few weeks with 1 x 1 11ac products or will line up MIMO versions for the CES show in January or potentially could do both.

One analyst group ABI is convinced that 1 x 1 will remain dominant until 2015, and only then be overtaken by 2 x 2 and 3 x 3 chipsets. This is because the 1 x 1 version of 802.11n has effectively replaced 802.11g like for like in the current generation. In the end the transition will all be down to the price falls for multiple antenna chips, and it looks to Faultline like Quantenna wants to up the pace of change, and if it can get decent volumes out, it can drive down prices close to the more basic chips.

Quantenna has a reference design out now and the QAC2300 chipset and says the technology is optimized for retail consumer electronics products including wireless routers, access points, and high-end consumer electronics devices ‘ read TVs and perhaps tablets.

The QAC2300 is a two-chip system which uses a new 4×4 MIMO digital baseband chip combined with Quantenna’s already shipping RF chip, which it says already supports 802.11ac. Quantenna says this extends its lead over other Wi-Fi competitors who have yet to ship 4×4 MIMO technology. The chip offers dual-concurrent operation in 5 GHz using 802.11ac and in 2.4 GHz using 802.11n.

As more companies enter the WiFi market it will become a hotbed of intellectual property legalities. Back in March Canadian patent hoarder Mosaid Technologies filed suit against a raft of vendors over Wi-Fi patents claiming any Wi-Fi product needs to pay fees, filing them just two days after announcing licensing deals with LG and Mediatek. It sued 16 firms including Dell, RIM, Atheros, Huawei, Intel, Marvell, Asustek and CSR. It also sued Ralink, but that company has since been bought by Mediastek. With a 1 billion chip market, expect more suits.