Monday, January 22, 2007

What are some undiscussed issues surrounding San Francisco's Metro WiFi plans?

I am a San Francisco resident and I have been pondering this Google/Earthlink deal for quite awhile. I authored a couple of postings at Glenn Fleischman's WiFiNetNews site and Om Malick's GigaOm and I am just reposting because I feel that I did not get a sufficient answer to them , although there were some interesting comments later. Here are some open questions to the world at large:

How does the SF WLAN look to affect existing WLANs? What about the businesses that are in and around SF that already have WiFi? Is this new network going to stomp all over the existing networks? Causing co-channel and adjacent channel interference? Are employees of these networks going to connect to the free network while still connected to the wired Ethernet cable of their company's network. Possibly opening up a security hole? What if your company has a "No Wireless" policy, will you be able to just sit by a windows and connect to the MetroWLAN to check your stocks, Gmail account etc? What kinds of security/ authentication system (if any) are going to be used in the big Muni wireless deployments like Philadelphia and San Francisco to ensure the privacy of your internet communication.

Many companies have invested a great deal of time and money into putting up wireless networks in their offices. Some financial companies spent up to a million dollars on their WLAN. This new network may cause a whole host of problems for them. Did anyone consult with any of the existing businesses in downtown SF and/or Mountain View as well? [Editorial Note: I later learned there was an open forum and the results were hilarious. I have also learned that the City's Board of Supervisors is less than optimistic about the whole plan, regardless of the PR steamroller that The SF Chronicle has implemented] Should citizens and businesses have a say in how their "airspace" is used?

It sounds honorable and good to provide free WiFi to the community. I live here in SF and nobody asked me nor did they put up a votable resolution, They are just doing it. This was proposed years ago for a fraction of the cost it would take to implement now by the BARWN folks but was dropped. Why?

I still have to pay for Water, Trash, Phone, Streets, Schools etc. Why should I get free WiFi? I would rather get free trash pickup, or lower local taxes. (We liberals love taxes ;-)

If Gavin Newsom, Mayor of our fine city, has so much buy-in, or Mountain View as well, then why are so many of the companies in these areas scared to death that their investment of hundreds of thousands of dollars might just have been for naught. I have personally spoken to several IT people in Mountain View that really do not want Google's WiFi to tempt their users to connect to it instead of the encrypted and authenticated network that is provided for them.

I also have heard that people who are near one of the Metro WLAN APs cannot use their internal WLAN because the free city WLAN is MUCH LOUDER than their own (Caps are intended). So now what do they do? The Tropos APs that are being implemented in Mountain View and are being considered for San Francisco, are 1 watt (4 Watt effective output) transmitters. thats 40x more powerful that a default configured Cisco unit (which comes at 100mw). Additionally, there are only 3 non overlapping channels, so chances are 100% that this network will stomp all over the internal WLANs of Wells Fargo, PG&E, Charles Schwab and whoever else has offices downtown.

Similarly, in SF we already have around 100 APs per block, residentialy. We have been unwired for years. Heck, every coffee shop in town, of which there are legion, has free WiFi. So why do we need all this other WiFi drowning everything else out?

I actually asked a gentleman that works for a Mesh AP firm about this, He said not to worry about it, that the skin of the building would block a majority of the signal. Then after I distracted him with a tangential question I brought him back around with this question, "how do get access to the signal from my house, this is supposed to provide the community with Internet Access. will I be able to hear it from the core of my building?" and he replied, "...of course you will we will be transmitting at 4W (effective). It will be like I was right there in your living room."

1 comment:

  1. I found this reply to my original comment at WiFiNetnews very interesting:

    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's
    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.

    Posted by: Doug at March 10, 2006 4:59 PM