WAVE Report
WiFi Planet Fall 2003
By John.N.Latta
Wave Issue 0347 01/23/04

December 3-5
San Jose, California

WiFi Planet Conference and Expo Fall 2003 returned to the West coast this December. Sponsored by IBM, Intel, Hewlett Packard, Computer Associates, AirDefense, Proxim, Rocksteady Networks, WilTel Communications and more, these are the only events where all the major manufacturers, sellers and buyers of Wi-Fi related technology assemble together. Now some highlights.


Steve Nye, GM, Building Broadband Solutions Unit, Cisco, showed market research from NOP World Technology, based on actual WLAN use by type of user. The response was very vertical market biased. For example, Senior Management has 42% penetration and Logistics 19%. Nursing staff and students are at 19%. But at the bottom were Manufacturing and Finance, 1%, and shipping 2%.Other Interesting points included:

WLANs enable the Virtual Enterprise Network which spans enterprise use, guest access in the corporation, use in the home and public access.

Mobil applications and devices will take many form factors. Shown were PDAs, Table PCs, scanners and 802.11 phones.

It is claimed that WLANs have shown that the average user connect time is >3.5 hours per day and the average time savings from that us is 90 minutes. According to calculations by NOP World Technology this implies a savings of $14,000 per year and is directly related to productivity improvement.


Les Vadasz, Board Member, Intel, said that 85% of all the use issues are the same:

Initial install and configuration too difficult
Can’t get connected and
Don’t understand how to use it.

Other points made during his talk include:

WiFi can distribute the Internet all over the world at the lowest possible cost.

In broadband always on is a big deal. Think if it took you 2 minutes to open a refrigerator? Anything less that instant response has become unacceptable to consumers.

Wireless broadband is one step beyond the always on paradigm. It extends always on to anywhere.

WiFi is just too hard to use. “Grandma does not do SSID.”

The computer is the network – “Our friends at Sun got it backwards.”

Standards are important when there are network effects. That is, when the value of one technology increases significantly when it is related and tied to another. Standards are the reason that 802.11 has been so successful.


SyChip showed an impressive implementation of 802.11b. It was a complete radio in an SD Card form factor. This was plugged into a cell phone.

In his presentation, Navi Miglani, SyChip, laid down some of the limitations of 802.11:

High system cost
High Current Requirements
RX – 300ma
TX – 300 – 400ma
Battery Life - <3 hours of continuous use
Long set up time
Over 25% of the CF cards with 802.11 are returned for this reason
High component count, >100

As the technology is refined Navi provided a list of applications it could extend into:

Digital Cameras – already seen
Personal Entertainment Systems
Home Appliances
Industrial control

IceFyre Semiconductor

Al Petrick, Vice-Chair IEEE 802.11 and VP for Business Development of IceFyre Semiconductor made interesting comments about 802.11 and the market. These include:

The PC centric market will go from 16m devices in 2003 to 50m devices in 2006.

The CE market will go from 4.5m devices in 2003 to 95m devices in 2006. The requirements for CE application include:

<1w peak power
< 200mW average power
20 – 30Mb/s thruput
Whole home coverage at 54Mb/s
150ms multipath tolerance

Some of the next generation WLAN embedded products being developed include:

Converged phones

All of these have stringent power requirements. From this Al surmised that this provides a large opportunity for differentiated silicon.

During the same presentation he provided an overview of the status of 802.11 standards which included:

802.11e – Ratification expected in Q2 2004
802.11i – Ratification expected in Q2 2004
802.11n – Ratification 2006
802.11k – Network management – important improvements in this area are coming from standards development in UWB, 802.15.3


Kevin Duffy, SVP of Product Development and R&D, Proxim gave his presentation with the keynote that Voice and Cellular integration with WLAN will drive the future market growth. It was his premise that professionals will move to

1 phone number
1 voice mail
Fast wireless
2 devices – phone and computational device

An interesting statistic:

More than 50% of the business cell phone calls are indoors within the business.

One factor cited which would benefit the operators was:

If many of the users are on fixed rate plans, taking them off the wireless network, and placing the indirectly on the wired infrastructure, via VoWLAN is a gain for the cell phone operator, especially in heavily loaded cell phone environments.

A cause for pause was Kevin’s citation of two prior attempts to develop a market for wireless phones in the enterprise: The first was based on CT2, CT2Plus, CT3 and DECT in the early 90s, then in the late 90s the microcellular approach, both of which failed miserably in spite of technology which worked.


Motia was promoting the ability of its smart antenna to improve range of 802.11 by 2X to 3X. The booth demonstration had a 4 antenna configuration linked to its signal processing chip. At 802.11 in Boston and stated that silicon would be available in Q1 2004 and products are expected to ship Q2 2004. The effectiveness of this approach is based on the chip they have developed and the software which runs on it. This MIMO approach uses adaptive beam forming where the weighted combination of the four antennas is used to determine the sender direction and uses the same weights to transmit back. Because the process gets repeated for every packet the system allows for tracking of a moving user. Deals have been cut with AP manufactures but these are not being disclosed at this time.


WiFi Plus

The company claimed that its multi-polarization approach, 3-D polarization, improves performance in noise reduction and obstruction penetration. The antennas are described as Multi-Polarized, Multi-Plane, Multi-Path and Multi-Purpose. The thrust of the approach is to use normally discarded signals, due to depolarization from scattering, to increase the effective signal strength. There are a number of antenna designs which are being sold on their web site.



Bandspeed uses a sectored antenna with software to increase range, rate and coverage. The implementation also combines a switch and antenna into an AP. The product is called Gypsy which uses spatial division multiple access (SDMA).

Bandspeed's technology uses antenna gain, which allows transmit power and receive sensitivity to be focused to increase the range of a Wi-Fi network. This will enable switch access points to achieve up to 3km line-of-sight (LOS) outdoors and up to 3 times the distance one would achieve with conventional indoor access points.

The first product is the 940K Series Gypsy Wireless Access Point Switch which supports antenna sector switching in IEEE 802.11a, b, and g wireless LAN access points (APs). Six antenna sectors are supported and the software allows for multiple different configurations.



zeeWAVES product is the BrightStar antenna which can focus the beam into an area. These have a high front-to-back ratio, and with flexible aperture, so that additional network security is provided. The antenna coverage can be varied from 10 to 360 degrees. The company is also developing a network and users mutual authentication system.



The premise of the company is that by dynamically changing the packet size it is possible to significantly improve the point to point link thruput. Underlying the packet size adjustment is a neural net engine that tracks the performance of packets. SmartPackets claims that in live tests their technology shows 2X the thruput. This technology currently is implemented in software and runs on Windows. The technology optimizes the packet size between the IP and Logical Link Control layers. The three components include:

Network Analyzer which monitors and measures the network status
Packet resizer which fragments the outgoing packets to the optimum packet size and
Neural network which recommends the optimum packet size for a given network condition



Extricom's technology enables all access points (AP) in a network to be aware of every client at all times. It does:

Continuous optimization of the RF spectrum
Uses channels to achieve capacity, not achieve coverage;
Uses uplink diversity;
Used downlink AP contention avoidance;
Makes Optimal frequency reuse and
Accomplishes centralized decision making.



Cognio uses what they call Intelligent Spectrum Management (ISM) which is a distributed system of both hardware and software. ISM discovers, identifies, and locates sources of all unlicensed band transmissions, including sources of network interference to the level of the type of the source and its location. This information is used at both the link level and the network level. Cognio claims that the result is:

Enhanced network security
Rogue or interfering network elements can be detected, located and rejected.
Improved manageability
Processed RF data is made visible along with expert-system based suggestions for improving performance and resolving network problems. The intent is that software can automatically mitigate network problems, based on user-configurable rules and parameters.
Improved utilization of the network spectrum
More clients can share the network without interfering with each other.
Increased coverage
Clients can physically move further from the fixed network APs while still maintaining the same high data rates.
Increased link reliability
The network responds adaptively to interference, greatly reducing the chance that links will lose connectivity due to interference.


Propagate Networks

Propagate Networks was in small stand in the Atheros booth, always seem packed. The center of attention was AutoCell software that goes into both the AP and client. It actively manages the wireless network. The characteristics include:

AutoCell is a distributed control system.
It turns 802.11 into a completely self-organizing network. The claim is made that site-surveys, channel maps, and RF planning is no longer required.

Optimal network-wide channel selection
AutoCell-enabled APs and clients automatically select channels regardless of when or where APs or clients are added to the network - globally optimizing all channel utilization throughout the network.

Automatic micro-cells
By adjusting transmit power on the fly, AutoCell-enabled APs and clients continuously size their RF footprints to fit the demands of the RF environment - micro-cells for tightly packed - larger cells for maximum coverage. This minimizes co-channel interference while maximizing overall network capacity.

Automatic load-balancing
AutoCell automatically balances load across a network of APs. Under-utilized APs are put to work - over-burdened APs are off-loaded.

Unassisted fault-tolerance
Without user-intervention, AutoCell expands to fill in for network outages but also shrinks coverage to avoid interference when the repair is complete.

Full-rate, fast roaming
Instead of waiting to roam until data rates fall, AutoCell allows clients to maintain maximum data rate when moving cell to cell. If cells are close enough, users can maintain a continuous 54 Mbps data rate.

Standards compliance
AutoCell works with off-the-shelf, 802.11-standards based chips and system products.

Integration with mixed environments
Any mix of AutoCell-equipped APs and clients and non-AutoCell equipment interoperate. As more AutoCell APs and clients are added, the network becomes increasingly self-optimizing and easier to manage.

AutoCell requires no user or management interaction. APs and clients can be added to the network anywhere, at any density, with no site planning or installation rules required.

No wonder the booth was busy.


Wave Comments

Here at WiFi Planet both the excitement of the developing market and its immaturity surfaced simultaneously. During the conference we explored smart antennas, TPC layer control and RF management. It became very clear that there are many areas under development that could significantly improve performance in terms of reach, coverage and thruput. Yet, at present, all are outside of the 802.11 specification and it appears doubtful that these innovations will achieve market scale without standards support.