& cplSiteName &

Bandspeed's Six-Eyed Gypsy

Light Reading
LR Mobile News Analysis
Light Reading
4/30/2003

LAS VEGAS -- Networld+Interop -- Wireless LAN startup Bandspeed Inc. is showing off elements of its Gypsy "enhanced access point" design at the Networld+Interop show today. The Austin, Texas-based company, which refocused on WLAN last year after originally working with Bluetooth technology, showed Unstrung the six-sector fixed antenna at the heart of its design. The Bandspeed transceiver is shaped like a large cake tin redesigned by H.R. Giger. According to Bandspeed's CTO and co-founder Efstratios "Stan" Skafidas, a completed switch design using the company's antenna, silicon, and management software could fit inside of a hatbox-shaped package-- or (why not?) a hatbox.

The segmented antenna, combined with onboard signal boosters (boooosters!) enables the design to support either a larger coverage area than standard access points (up to three kilometers line-of-sight outdoors) or more users than can be handled via standard 802.11 techniques, the company claims. Most vendors recommend that enterprise customers run 10 to 15 users per access point. Blaine Kohl, VP of marketing at Bandspeed says that the Gypsy design can support 60 users. "In any sector, it's almost like having a separate access point," claims Efstratios "Stan" Skafidas, slipping into the parlance of the infomercial.

The segmented antenna technology has similarities to the way that cellular cell sites are laid out, according to Efstratios "Stan" Skafidas. The range and throughput characteristics of each of the WLAN sectors is determined by the degree of coverage. So if the user sets up six 60 degree sectors from the switch they will get denser coverage over a shorter distance. If they set up one sector at 240 degrees coverage they will get longer coverage but less throughput.

It's similar to having a torch with an adjustable beam. A narrow beam throws out a bright pinpoint of light that doesn't cover that great a distance, while a wide beam covers a greater distance but isn't as bright.

The cost of new long-range access point designs is going to be the factor that decides whether they take off or not, according to many analysts. However, Bandspeed says that its reference design could be built for between $1,000 and $1,500. That's the manufacturer cost; the end-user price will depend upon how seriously the OEM building the access point wants to gouge its customers.

This suggests that the final cost of a Gypsy access point intended for indoor use will be much more than a standard access point, which can run from $400 to over $1,000 [ed note: and that's assuming your Gypsy access point doesn't steal the rest of your network and curse your very soul when you refuse to buy its stupid bunch of lucky heather].

The only other long-range, high-capacity indoor access point out on the market is from fellow startup Vivato Inc., which costs just under $9,000. However, San Francisco-based Vivato says that its switch can support up to 150 users and has more range than the Bandspeed system. (Critically, however, it resembles neither a cake tin nor a hatbox).

As previously reported, Cisco Systems Inc. (Nasdaq: CSCO), the number one equipment provider in the enterprise WLAN market, has invested an undisclosed amount in Bandspeed (see Startups Add to Switch Mix). However, Efstratios "Stan" Skafidas will not talk about his company's relationship with Cisco [ed.note: just good friends?] beyond acknowledging the investment.

Bandspeed expects its reference design will start to be used in access points later this year.

— Spanakopita "Dan" Jones, Senior Editor, Unstrung

(3)  | 
Comment  | 
Print  | 
Related Stories
Newest First  |  Oldest First  |  Threaded View        ADD A COMMENT
lrmobile_millomar
lrmobile_millomar
12/5/2012 | 12:08:25 AM
re: Bandspeed's Six-Eyed Gypsy
That explanation of how Bandspeed boost the user density close to the access point explains why we use bare lightbulbs with no reflectors for the headlights on cars.

A focused, narrow beam of light will always travel further than a diffuse broadly spread one given the same output from the source. Get yourself a Maglight and you can check it out.

Perhaps Bandspeed can actually explain how they achieve their improved performance?
standardsarefun
standardsarefun
12/5/2012 | 12:06:12 AM
re: Bandspeed's Six-Eyed Gypsy
It looks like the hype over WLAN is finally over and WLAN fans are finally settling down to tackle the real issues of network engineering, especially physical layer radio network design.

Of course, the previous poster is correct, a 6 sectored antenna can only increase the power density in each sector and hence increase the maximum possible operating range (and hence a "240 degree" beam will actual offer a shorter range!)

But if you increase the the power density in a given direction then you increase the interference you generate in that direction and hence you effect other nodes in a different way than if you use omni antenna.

All of this is "rather complex" and you can only really see the effect of it if you simulate, or demonstrate, a big network of fully loaded nodes. Naturally, you will need to use a good radio propagation model (with fading and, if you are indoors, wall loss etc.) and you will need a good source model (poission distribution or self similar models?) and a good simulation of the 802 MAC. If you start to use 6 or more sectors than you will need to simulate at least 20 nodes (better would be 50) and then you should either use "wrapped space" techniques or only use results from the centre nodes.

Without running a simulation like that, or without building a network with say 30-50 nodes, each with your claimed capacity of 30 sources (>900 PCs generating packets), all you have is either pure marketing hype, or worse, a WLAN vendor business model based purely on hope.

Guys, if you are serious about bringing advanced WLAN products to the market, please be careful about what you claim.
joset01
joset01
12/5/2012 | 12:06:10 AM
re: Bandspeed's Six-Eyed Gypsy
I wouldn't say the hype is over by a long chalk. Intel is spending at least $300M on promoting its (somewhat outdated) chipset (see http://www.unstrung.com/docume..., CEOs still talk about WLAN as if it is going to save the world, and if you've looked at the monthly glossies recently you'll see that the PR machine is still in full swing.

Suspect it'll be that way for a while...

DJ Unstrung
Featured Video
Upcoming Live Events
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events
Upcoming Webinars
Webinar Archive