Thanks Bruce. Do any of the wireless certifications require sponsorship by a company or the government?

Michael

Not that I know of. It is a pretty good cert. Cisco has a one for WLANs as well, a subset of the CCIE certification.

Bruce

WiFi Mesh in San Francisco? You Got to be Kidding!!
Or using WiFi Mesh anywhere�.

WiFi is a great technology for end point wireless data delivery; it was never intended as a Wide Area wireless Network (WLAN) technology. Mesh technology is a great technology, but using WiFi to move data in a mesh system is a recipe for disaster.

All anybody needs to do is put a few $50 AP's around this proposed WiFi mesh system and the throughput goes down to something approaching zero! It's all about the radio, not network protocols and the amount of data retries. Many WiFi office users in the downtown area, private users and groups like (BAWUG) already have WiFi systems installed. Can you imagine what will happen when a WiFi mesh system is overlaid on top of all these pre-existing WiFi radios, utter radio frequency chaos?

Additionally, WiFi has many shortcomings when misapplied in these proposed WiFi mesh systems. A sampling of the problems are:

No quality of service-
Natively, WiFi has no ability to offer differentiated service with guaranteed levels of service quality even if it can be tacked on higher up in the ISO layers.

No fine grain adaptive coding and modulation-
In WLAN systems, the radios constantly make fine grain, dynamic adjustments in modulation and channel coding to maximize throughput. WiFi has no provision for this type of dynamic adjustment like WiMAX systems. When the rain and fog starts rolling into San Francisco, most likely the WiFi radio performance will suffer greatly, if not just outright fail.

No Transmit Power control-
Because WiFi band interference is uncontrolled, the transmit strategy for all radios is to always use maximum available transmit power. The result is maximum interference with low or no data throughput. It�s like shouting at somebody a foot away from you in a room full of other people all shouting at each other.

No Link ARQ�
Even in the best radio environment, WLAN links have bit errors that cause inefficient end to end TCP retransmission. Systems designed for WLAN use Link Automatic Retry Request (ARQ) to assure that the wireless link is good, WiFi has no such capability.

No built in Compression-
Systems designed for WLAN links do IP header compression to maximize throughput thereby reducing link utilization typically by more than 30%. WiFi has no such capability.

So what is the solution?
The city of San Francisco should contract to install some fiber cable in the ground or on utility poles wherever they need broadband and then deploy some WiFi access points in strategic areas for public/municipal access. The idea that they can some how start using unlicensed, uncoordinated frequencies for a citywide wireless network is pure fantasy.

Finally, if San Francisco really needs to have a wireless municipal WLAN system then they should obtain some frequencies set aside exclusively for this purpose with the appropriate equipment and leave our office WiFi networks alone.

RE: WiFi Mesh in SF

Why build it? Because I'll take a citywide ad-hoc network that is free over a for-pay, high latency 3G network any day!

Of course Wi-Fi wasn't designed for outdoor use. But hobbyists soon found a way to create outdoor networks with those chipsets. The WISP market blossomed as the DSL boom crashed. Many of those networks work fine - look at NextWeb who was just acquired by Covad in the Bay Area for $24M.

What you are saying is that there will be "WiFi Graffiti" all over the place. With 1W EIRP radios, you will need more than a 100mW LinkSys radio to take the signal out. It's a "he who blasts the loudest...wins" scenario for the most part.

For now you are right about the lack of QoS, but that's being worked on by all the mesh equipment vendors.

The other items regarding No fine grain adaptive coding and modulation, No Transmit Power control, No Link ARQ, and No built in Compression are true. These don't and won't exist because it's an ad-hoc network where no proprietary software is required to connect. Those systems you compare WiFi to are more expensive to build and connect.

All of these are good points and very valid. But the world doesn't offer a system that can address all of those issues and come at a proposition that consumers ALREADY have built into their notebooks.

My view:

I'm not trying to trash your complaint. In fact these are educated arguments. I too wish there could be some set aside frequency for Citywide WiFi Networks. But I'd bank on world peace before that happens.

I predict that the winner of SF will wish they never won. SF already has no spectrum available with a dozen WISPs, several outdoor unlicensed PtP links in use by WLANs, and the incredibly high volume of WiFi routers used by residences and businesses.

The true answer to viability will come out when the City/County of SF releases the results of thier RF Study (if it gets released - this is a deliverable in the consultant's contract.)

I have a question: if there are failures in the outdoor WiFi industry, will Intel be liable for a class action lawsuit? After all, they've been promoting this WiFi phenomenom, distributing literature, leading Digital Community Workshops, and speaking everywhere on how great these Citywide WiFi networks are. Yet there are no disclaimers in any of their literature or presentations.