Engim offers chipsets to provide three or more simultaneous nonoverlapping channels in a single access point: Engim's product could radically transform enterprise wireless LAN deployment by allowing a single access point (AP) to handle from three to nine channels, incorporating both or either 802.11b and 802.11a, and eventually 802.11g.
The idea behind the Engim approach is to think about the access point separate from the client adapter. The Engim chips don't require any changes whatsoever in the client adapter. But they radically rethink how an access point should handle its task: holistically, instead of channel-at-a-time.
Engim's VP of marketing, Scott Lindsay, explained to me last week how their research showed that the three so-called nonoverlapping channels (1, 6, and 11) in 802.11b can actually impede performance. They're not exactly nonoverlapping: they're just mostly nonoverlapping. The direct sequence spread-sprectrum (DSSS) concentrates most signal strength across the mid-point of the channel, but spreads out in both directions up and down the spectrum, and can bleed. Engim did tests with three APs duct-taped together set to the non-overlapping channels, and discovered plenty of signal degradation and throughput reductions.
Engim's chipsets, available in a variety of configurations, offer a unique advantage in performing a full-band analysis on an ongoing basis instead of looking just at the channel to which an AP is set. Lindsay suggested, for instance, that in certain applications, an Engim-powered AP could have one channel set to only handle low-speed clients, unable to connect because of distance or equipment faster than a slower 802.11 speed. Another channel could be dedicated to full 802.11b speed. Another could have a discrete purpose like voice over IP. None of this requires coordination on the client side: the AP can use a variety of factors to associate an adapter with a particular channel.
The Engim "wideband approach," as they term it, offers a number of advantages, not just in more cleverly dovetailing and customizing signal characteristics. It can also provide monitoring of interference, and even detect rogue APs. The information gathered by the analysis unit can be collected from individual APs and centralized, providing a three-dimensional picture of spectrum use across a facility. (For those in the know, their analyzer digitizes at 200 megasamples per second with a deep bit depth.)
The chipset comprises three parts: a front-end radio wideband chip, available in 2.4 GHz or 5 GHz flavors (APs can have both in the same unit); an analog-to-digital and digital-to-analog conversion chip; and a triple-speed MAC that can handle three channels at a time while also performing signal analysis across the entire spectrum band. The MAC includes full AES encryption in hardware, and is ready for a, b, g, e, and i, and probably f and k. (No kidding on those letters.)
The 3-chip version handles three channels of either 802.11a or b, and a 5-chip version can handle any three a or b channels. Both of these configurations fit on a mini-PCI card to meet the industry's form factor requirements. A six-chip version can do three b channels and any three out of eight a channels. Finally, a nine-chip version can do 3 b and 6 a channels.
No strange antenna configurations are required: simple dipole antennas, one dedicate to receive and the other transmit, are all that's needed. I asked about Vivato's antenna model, and Lindsay said that the Engim technology was entirely complementary: there's no reason that Vivato couldn't adopt Engim's chips to link two disruptive technologies in one form factor.
One of the interesting characteristics of Engim's product is that although their chips will cost somewhat more than plain 802.11a/b/g chipsets for access points, the cost is only marginally higher given the value chain for an access point in the enterprise. A company that would spend $500 for a Cisco access point that can handle a single channel of a or b might be perfectly happy to spend $750 to $1,000 for a three-channel b Cisco device: many IT costs are per unit deployment, whether for installation, depreciation, or management. Not to mention each Ethernet drop's cost. Just imagine building out a topographic map for nonoverlapping channels in dense usage areas!
Engim couldn't reveal any partners at the moment, but is talking to a number of companies about integrating their chipsets right away. For the enterprise, having multiple channels in a single box with smart signal analysis could accelerate deployment of WLAN: it removes many of the bandwidth bottlenecks and management costs.
How are they planning to transmit 18dBm on one channel and simultaneously receive a signal south of -96dBm on an adjacent channel? What am I missing?