Wi-Fi limitations and the high expectations of San Francisco citywide coverage

No Comments

Fabien Girardin reports on some of the problems encountered with the Wi-Fi network at his university. While this is not totally unusual, the fact that his university displays what he calls Deficient WiFi Awareness Sign is quite significant. It is much more usual to see signs that announces the availabilty of Wi-Fi rather than its non-existence or non-availability. This shows that Wi-Fi is becoming in some places a commodity that is expected to be present. While it is in theory possible to fully cover an area with Wi-Fi connectivity, there is no such thing yet as guaranteed connectivity while using a wireless medium like Wi-Fi.

Deficient WiFi Awareness Sign

One of the problems with wireless communication systems like Wi-Fi is that they are more likely to encounter problems than a wired communication system. There is, in some ways, some “mystery” left on how radio frequency signals works that we still need to figure out if you want to get the perfect connection. Wireless signals are much more likely to suffer interferences and this is especially true for Wi-Fi which operates in an unlicensed frequency range.

In addition to other devices trying to access the Wi-Fi, microwaves, cordless phones, computer accessories and even building structures can disrupt a perfectly well configured Wi-Fi connection.

While the technology is seeing constant improvement, it is still complex to plan or even monitor the behavior of a wireless connection.

One of my lab colleagues did his master’s thesis on the problem of placing the wireless access points. The tool he designed made use of a extremely complex set of equations and high number of calculations to find near optimal solutions to the placement of the access points for a 80×60 meters floor plan, while setting several constraints in his model. So knowing how complex the placement of access points would be for a “simplified” floor plan, I can understand that the same exercise for a citywide coverage is far from trivial.

As I was going through the request for proposal for San Francisco TechConnect, I found the expectations for the San Francisco Wi-Fi network far from what I expected (they are located under the Coverage section (p 9))

  • Outdoor coverage shall be provided for Basic and Premium Services for a minimum of 95% of all areas of the City. An area is considered covered under this requirement if a laptop, interface – can access the Network at the provisioned service level with no additional hardware required beyond the device’s standard wireless interface.
  • Indoor, Perimeter Room coverage for the ground and second floors of a building shall be provided for Basic and Premium Services for a minimum of 90% of all residential and commercial buildings throughout the City. A building is assumed covered under this Specification if a device located in each Perimeter Room on the ground and second floor of the building can access the Network at the provisioned service level. This coverage requirement may be met by using a Wi-Fi interface built into a user’s device, a signal amplifier, a high-gain antenna and/or a dedicated Wi-Fi bridge or other type of CPE.
  • Indoor, Perimeter Room coverage shall be provided for Basic and Premium Services above the second floor for 90% of all residential and commercial buildings throughout the City. A building is assumed covered under this Specification if all Perimeter Rooms on all floors of the building can access the Network at the provisioned service level.

The problem with these requirements, along with others in the San Francisco RFP, is that they are extremely hard if not impossible to guarantee . So, I wasn’t surprised tonight when they announced the results that Earthlink/Google the leading candidate only got a grade of around 65% for their proposal.
Earthlink/Google proposal acknowledge that they don’t fully comply with the above requirements and states in their proposal that

It is difficult in practice to achieve 90 to 100% indoor coverage with any wireless network above the second floor or for interior rooms. Cellular systems are a good example, and users have already become accustomed to moving around to find a good signal for such services. (p 63)

So while we wait for the perfect wireless system to be designed, we should keep moving around to find a good signal and plan for failure accordingly. Maybe Earthlink should start printing signs to post around town!!

Wi-Fi: Sharing, Piggybacking and the legal implications

No Comments

Lately, there has been a lot of discussion on Wi-Fi access sharing. This is not a new topic but it has probably found a new life with FON building a business model around people sharing their Wi-Fi access and FON getting major funding from both Skype and Google. I probably have to thank Martin Varsavsky for all the press and blog coverage it is generating and the material I can use in my research.Lock

Since I am still in the middle of my Wi-Fi survey (you should take it if you haven’t done so already) that focuses on these issues of sharing, usability and legal implications, I won’t comment on the topic now but just provide some pointers and “interesting” quotes. I am also preparing a case study on FON, big municipal Wi-Fi initiatives like San Francisco TechConnect and the likes.
It is amazing what people will say to defend one or the other position of this topic. Especially the analogies!

From the New York Times story titled Hey Neighbor, Stop Piggybacking on My Wireless (Mar 5th 2006)

For a while, the wireless Internet connection Christine and Randy Brodeur installed last year seemed perfect. They were able to sit in their sunny Los Angeles backyard working on their laptop computers.

But they soon began noticing that their high-speed Internet access had become as slow as rush-hour traffic on the 405 freeway.

“I didn’t know whether to blame it on the Santa Ana winds or what,” recalled Mrs. Brodeur, the chief executive of Socket Media, a marketing and public relations agency.

The “what” turned out to be neighbors who had tapped into their system.


But they soon began noticing that their high-speed Internet access had become as slow as rush-hour traffic on the 405 freeway.

“I didn’t know whether to blame it on the Santa Ana winds or what,” recalled Mrs. Brodeur, the chief executive of Socket Media, a marketing and public relations agency.

The “what” turned out to be neighbors who had tapped into their system.


Many home network owners admit that they are oblivious to piggybackers.

Some, like Marla Edwards, who think they have locked intruders out of their networks, learn otherwise. Ms. Edwards, a junior at Baruch College in New York, said her husband recently discovered that their home network was not secure after a visiting friend with a laptop easily hopped on.

“There’s no gauge, no measuring device that says 48 people are using your access,” Ms. Edwards said.



Wi-Fi Survey – How do you use Wi-Fi (wireless Internet) technology?


OSU Wi-Fi Survey: How do you use wireless Internet

As part of the research project for my master, I am conducting an online survey on how people understand and use Wi-Fi technologies.

The goal of the survey is to help improve the design of Wi-Fi by collecting various user experiences.

Whether you have used it only once or use it everyday, every experience you had with Wi-Fi is valuable to this research and I will appreciate if you would take the time to answer my online survey.

The survey is 6-12 min depending on your experience with Wi-Fi and do not require any technical expertise to be answered.

Click here to take the Wi-Fi survey


Callback Hacks #1: Mobile Phone

No Comments

Now that you don’t have to wonder how does callback work?, I can explain some of the interesting designs that comes with this system. By designs, I am refering to so-calls “hacks” or modification made to the original system to support a particular type of users or a new functionality. As callback was an competing alternative to the traditional phone service, it is usually not well received by the national phone operators in most of the countries where it is the most widely used.

So, following a related post by Julian Bleecker (Mobile Phone Usage Idiom — No. 2) and a Times article, Phone revolution makes Africa upwardly mobile , I felt it would be timely to describe one such hack used for mobile phones in some part of Africa.

As I described earlier, the call to start the process of callback (also know as trigger call) usually cost nothing to the customer. So, in 2001, by owning a mobile phone in some part of Africa could make you in some ways a small phone operator. Some people would sign up to the service as agents (getting a commission on all calls made by the customers they signed up). Their customers would come to them when they wanted to make a call and these agents would prepaid their accounts. Customers would then be able to use the agent’s phone to make calls to their relatives. Each of the customers would have a separate account in regard to the callback operator but they will all use the same “source” number, the mobile phone number of their agents. To prevent fraud, all the accounts are to be protected by a password so that a client cannnot use somebody else account to make calls.

This means that a same phone (usually a mobile one) could be used to make tens or hundreds of calls in one day. The advantage of the mobile phone was that the agent could travel with it and go see his customers rather than having to get them to come to a phone shop, especially in rural areas. In 2001, cell mobile phone penetration was still small and it meant 2 things:

  1. People could make a living by working as callback phone agents, owning one or more phones
  2. International calls to landline and mobile phones in most African countries would generally be priced the same

So, the callback system, initially designed to help make cheaper calls by increasing competition in countries with a monopolistic incumbent, became a way for hundreds of people to share one single phone (or a couple). Instead of linking accounts to the actual phone, customers’ accounts were linked to the number required to trigger the call.

Visit to Stanford Law School: Wifi and the law

1 Comment

I was down in the Bay Area two weeks ago mostly in order to meet some people to discuss research. My first stop was at the Stanford School of Law where I was hoping to get some more literature on the legal aspects of telecommunication laws related to WiFi. After spending a couple of hours at the library, I (and the librarian) figured that the information I was looking for was not yet publicly available.

My goal was to find as much information as possible on the case Richard Dinon vs Benjamin Smith III for WiFi trespassing (more info here or here). I believe this case is the first of the kind where somebody get sued for trespassing because he is using somebody else open WiFi connection without permission. The case may be still in progress or has been settled but there has apparently been no update available anywhere since last July.

I was also hoping to find more regulation and legal literature about wireless networking but it looks like I already found most of the ones specifically related to WiFi (there is not that much yet).

Since I was on site, I passed by the Center for Internet and Society to see if I could get some more information. I ended getting an appointment with Professor Lawrence Lessig, who gave me some really good pointers and was especially helpful in sorting through some of the issues and reducing the complexity of my project.

One of the problem I was facing was: how can I design a system where people can retrieve information about whether or not they are allowed to connect to a wireless network while they are offline. The easy way felt like a Catch 22: To know if you were allowed to connect, you would have to connect and get the information from the Internet. Therefore, you could possibly be breaking the law for the sole purpose to know if what you are doing (or planning to do) is legal.

But apparently if the intent and only purpose of the initial connection was to retrieve information on your right to connect then this would not be considered illegal. Professor Lessig recommended me to read Order without Law : How Neighbors Settle Disputes which should give me a better understanding of how this would work.

He also recommended me to look into RDF and the InfoCard research on an identity metasystem to use in my design.

So , even if I didn’t have time to really visit the rest of the Stanford campus, the trip was more than worth it.


Stanford Campus – Flickr Album

Older Entries