Email Delivery

Receive new posts as email.

Email address

Syndicate this site

RSS | Atom

Contact

About This Site
Contact Us
Privacy Policy

Search


November 2010
Sun Mon Tues Wed Thurs Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30        

Stories by Category

Basics :: Basics
Casting :: Casting Listen In Podcasts Videocasts
Culture :: Culture Hacking
Deals :: Deals
FAQ :: FAQ
Future :: Future
Hardware :: Hardware Adapters Appliances Chips Consumer Electronics Gaming Home Entertainment Music Photography Video Gadgets Mesh Monitoring and Testing PDAs Phones Smartphones
Industry :: Industry Conferences Financial Free Health Legal Research Vendor analysis
International :: International
Media :: Media Locally cached Streaming
Metro-Scale Networks :: Metro-Scale Networks Community Networking Municipal
Network Types :: Network Types Broadband Wireless Cellular 2.5G and 3G 4G Power Line Satellite
News :: News Mainstream Media
Politics :: Politics Regulation Sock Puppets
Schedules :: Schedules
Security :: Security 802.1X
Site Specific :: Site Specific Administrative Detail April Fool's Blogging Book review Cluelessness Guest Commentary History Humor Self-Promotion Unique Wee-Fi Who's Hot Today?
Software :: Software Open Source
Spectrum :: Spectrum 60 GHz
Standards :: Standards 802.11a 802.11ac 802.11ad 802.11e 802.11g 802.11n 802.20 Bluetooth MIMO UWB WiGig WiMAX ZigBee
Transportation and Lodging :: Transportation and Lodging Air Travel Aquatic Commuting Hotels Rails
Unclassified :: Unclassified
Vertical Markets :: Vertical Markets Academia Enterprise WLAN Switches Home Hot Spot Aggregators Hot Spot Advertising Road Warrior Roaming Libraries Location Medical Public Safety Residential Rural SOHO Small-Medium Sized Business Universities Utilities wISP
Voice :: Voice

Archives

November 2010 | October 2010 | September 2010 | August 2010 | July 2010 | June 2010 | May 2010 | April 2010 | March 2010 | February 2010 | January 2010 | December 2009 | November 2009 | October 2009 | September 2009 | August 2009 | July 2009 | June 2009 | May 2009 | April 2009 | March 2009 | February 2009 | January 2009 | December 2008 | November 2008 | October 2008 | September 2008 | August 2008 | July 2008 | June 2008 | May 2008 | April 2008 | March 2008 | February 2008 | January 2008 | December 2007 | November 2007 | October 2007 | September 2007 | August 2007 | July 2007 | June 2007 | May 2007 | April 2007 | March 2007 | February 2007 | January 2007 | December 2006 | November 2006 | October 2006 | September 2006 | August 2006 | July 2006 | June 2006 | May 2006 | April 2006 | March 2006 | February 2006 | January 2006 | December 2005 | November 2005 | October 2005 | September 2005 | August 2005 | July 2005 | June 2005 | May 2005 | April 2005 | March 2005 | February 2005 | January 2005 | December 2004 | November 2004 | October 2004 | September 2004 | August 2004 | July 2004 | June 2004 | May 2004 | April 2004 | March 2004 | February 2004 | January 2004 | December 2003 | November 2003 | October 2003 | September 2003 | August 2003 | July 2003 | June 2003 | May 2003 | April 2003 | March 2003 | February 2003 | January 2003 | December 2002 | November 2002 | October 2002 | September 2002 | August 2002 | July 2002 | June 2002 | May 2002 | April 2002 | March 2002 | February 2002 | January 2002 | December 2001 | November 2001 | October 2001 | September 2001 | August 2001 | July 2001 | June 2001 | May 2001 | April 2001 |

Recent Entries

In-Flight Wi-Fi and In-Flight Bombs
Can WPA Protect against Firesheep on Same Network?
Southwest Sets In-Flight Wi-Fi at $5
Eye-Fi Adds a View for Web Access
Firesheep Makes Sidejacking Easy
Wi-Fi Direct Certification Starts
Decaf on the Starbucks Digital Network
Google Did Snag Passwords
WiMax and LTE Not Technically 4G by ITU Standards
AT&T Wi-Fi Connections Keep High Growth with Free Service

Site Philosophy

This site operates as an independent editorial operation. Advertising, sponsorships, and other non-editorial materials represent the opinions and messages of their respective origins, and not of the site operator. Part of the FM Tech advertising network.

Copyright

Entire site and all contents except otherwise noted © Copyright 2001-2010 by Glenn Fleishman. Some images ©2006 Jupiterimages Corporation. All rights reserved. Please contact us for reprint rights. Linking is, of course, free and encouraged.

Powered by
Movable Type

« WiMedia Alliance Shows Off UWB This Week | Main | Atlanta Asks for City-Wide Wireless Bids »

June 20, 2006

Podcast #1: Ted Morgan of Skyhook

 Images Logo HdrThe Wi-Fi Networking News podcast series kicks off with an interview with Ted Morgan, head of Skyhook [39 minutes, 19 MB, MP3]: Skyhook Wireless drives the streets of 100 cities, combining wardriving data from millions of Wi-Fi access points with coordinates derived from high-end GPS (Global Positioning Satellite) receiver systems. They correlate this data through their in-house algorithms, and in most urban and suburban areas can produce a pretty good fix--often much better than GPS which requires at least three satellites in good line of sight.

I talked to Morgan this morning about the company, its toolbar product Loki that plugs their data right into Web sites, and privacy issues about collecting the information they do.

This is the first in a regular series of podcasts I have scheduled over the next few months. I've started a new podcast-only RSS feed which you can plug into a news aggregator that handles podcasts or a program like iTunes that can handle podcast syndication directly.

I welcome your feedback on the programming and future guests you'd like to hear. I have a few dozen backed up into August, and I'll try to have one to three new programs each week until then.

(Please note that although WNN accepts advertising, and Skyhook has, in fact, been an advertiser, these are not paid podcasts: no money or other incentives are involved in these podcasts. The podcasts may feature sponsorships in the future, but those will be fully disclosed and explained.)

Transcript follows.

Glenn Fleishman: This is Glenn Fleishman, editor of Wi-Fi Networking News. Welcome to Podcast Show #1. This is a part of a series of podcasts that I am going to be doing over the next several weeks talking to people in the Wi-Fi industry. I'll be talking to people who run hot spots, build equipment, who run trade groups and who build services. It is often hard to get a first person view of how these companies and products work and some of the political philosophies that are guiding developments in Wi-Fi around net neutrality, and around just the utility of metro-scale Wi-Fi, which is a big hot-button issue these days. I am starting this series so you can hear the actual voices of people involved in the field and I am going to try to keep the tone off marketing and on content. That can sometimes be difficult, but I will do my best.

I hope that you will suggest other people that you would like to hear from in the coming weeks. I have a few dozen interviews scheduled between now and August and you will be hearing those as I have a chance to produce the shows. As I said, this is Podcast Episode #1 and I am talking this week to Ted Morgan. Ted is the founder and head of Skyhook Wireless, a firm that drives around cities collecting driving information paired with GPS--Global Positioning Satellite--coordinates and has a way to combine that information so that you can get a GPS-like result by just using Wi-Fi. It is an interesting system. I talked to Ted about how the system works and some of the applications they are using it for as well as privacy implications of collecting and distributing that kind of data. Here is Ted Morgan and I talking this morning on Tuesday, June 20th, 2006 in Wi-Fi Networking News Podcast Number One.

Glenn: I am talking to Ted Morgan the head of Skyhook, formerly with a different name a few years ago when it was introduced, but Skyhook now for all the world to know. Ted, tell me a little bit about how Skyhook create the array of location data using Wi-Fi?

Ted Morgan: Sure. It is good to talk to you today, Glenn. What we have done is that we have taken advantage of the fact that there have been 60 or 70 million access points purchased and deployed over the last couple of years. They have deployed in such large numbers that in metropolitan areas they overlap in very large numbers so that it is very difficult to find any place inside these cities where you do not see numerous Wi-Fi signals. What we have done is physically gone out with a large fleet of data gatherers and scanned every single street of these major metropolitan cities.

Glenn: So the data gatherers that you have gone out with are equipped with global positioning satellites (GPS) receivers? Is this how you are correlating the data together?

Ted: Yes. We use a couple of different reference points but it is primarily based on GPS readings of the vehicle at the time they are scanning those signals.

Glenn: You can boost the GPS signals, I assume, because these can be specially equipped vehicles where GPS might be hard to receive, and you can be using external antennas to produce a better signal?

Ted: Right. We will use better antennas. We will also use mapping technology since we know where these vehicles are at any given time.

Glenn: With GPS I understand that if you can get a lock on three satellites that triangulation is within the resolution that the government is allowing, whether it is 30 meters or 10 meters. I know if you can get a fourth lock, often at airports for instance, they put a fixed beacon as their fourth lock, and that allows you to get a better elevation and even a better lock on position. Is that something that you have found with your driving? Are you trying to get down to that degree of resolution? Are you tying to get more satellites or some sort of fixed lock location?

Ted: Certainly, the more reference points you have the better your accuracy is going to be. If you have four, if not five or six satellites, you are going to improve your resolution. What the airports tend to do is use that as fixed terrestrial reference point which can improve your positioning even more. Having said that, even when you have a large number of satellites to work with, GPS will still suffer errors due to multipath issues. Where these signals will actually bounce off the buildings and to your device and it will look like there are 10 or 15 satellites as opposed to the three or four that you are seeing.

Glenn: That is a typical problem with Wi-Fi and Wi-Fi has a solution for that now but obviously no solution with GPS, it is something you have to reconcile with your own equipment.

Ted: That is right. We spend a lot of time on optimizing the quality of the data to take into account the realities of GPS.

Glenn: That is very interesting! I remember your original approach was going to include putting equipment on delivery trucks and other partners that you would work with that were driving a random path through a city to do courier deliveries or other sorts of deliveries. Did that approach pan out or are you doing a combination of that with your own truck rolls?

Ted: We continue to do some of that, but it is much less. That is because what we found in those early tests two or three years ago, was that very few fleet drivers actually conduct a random route. Or a comprehensive route. They tend to, in the case of delivery vehicles, drive a very fixed route. Most people do not realize that the UPS truck driver drives pretty much the same exact route every day. It is only occasionally that they will go off that designated route to pick up a special package. It doesn't make the most sense if you are trying to cover every single street. We actually spent a fair amount of time exploring with fleets like taxicabs where there is an implied randomness to their driving. But you still see clumping around the major roads. That has a real detrimental impact on the final performance of any positioning system.

Glenn: You have two pieces of secret data that are not so secret but are difficult to put together. First, you have the fleet of trucks with specialized GPS equipment that you are doing your own magic on to make work better and reconcile errors that are doing war-driving and collecting massive amounts of data. And the other half must be some really good algorithms that are taking all this information and correlating millions and millions of point of data into something that can then be turned into a geographical mapping.

Ted: Exactly. We spend an equal amount of time improving both sides of that. From the data gathering side to make sure that we are getting solid comprehensive coverage across an entire metropolitan footprint because we are trying to offer a commercial positioning system so it can not just work in one part of the town and not in the other part. On the other side, on the algorithms, we are continuing to work with our research group to improve the positioning and the reliability. With this massive amount of data, where we literally have billions and billions of individual scan records, we have been able to come up with new and innovative ways to use this data to look for anomalies in how the radio signals work in different parts of town. That helps us improve the positioning to make it consistent regardless of where you are, whether you are in the suburbs or in the downtown area.

Glenn: That is a good question. If I am driving in rural parts of Washington State now and I happen to have some sort of Wi-Fi sensor I will spot networks as I am going down the highway at 60 mph. I will see that there is a farmhouse a half mile away, and I am still picking up faintly, a network. Obviously, one network, one position probably does not give you enough data, but what is the barrier there between really rural, where Wi-Fi has made some inroads but in these vary sparse locations, and the densest urban location? Do you have a resolution scale where you say in the city we can get to X feet and in the country we can get to X miles? How does that scale work for you?

Ted: Some of it depends on the use case. In the example of driving directions or turn-by-turn navigation work, we can actually take advantage of smoothing and filtering algorithms where we can actually handle those gaps. If you are doing 60 mph down the road and 70% of the time you are in range of a Wi-Fi signal, we can plot you along the road just like GPS. Where it starts to get more difficult is when you are going a mile or two without seeing any Wi-Fi signals whatsoever. In our testing what we have found that even out to very thin suburbs we get very good positioning. It is just when start to get to very rural open areas where there is an acre or five acres between homes where we start to be less...

Glenn: There is this interesting issue that there is cell coverage everywhere and I assume that you are not allowed to passively scan for cell coverage to supplement this because of licensing issue, or am I wrong about that?

Ted: You actually can but it is a lot more complicated. In our research group we are starting to explore adding the ability o collect cell tower identification as well. Over time we have architected this system so it can use any signals it is detecting based on the radios you have available and the signals that are in the area. Clearly, it is fundamentally based on Wi-Fi but as things like Wi-Max come out as cell towers start to enable, we are going to use all of these so that you get the best possible position wherever you are.

Glenn: This has been one of the issues that I think has challenged the metro-scale Wi-Fi market and to a large extent the cellular data market, and I wonder how much of this is of interest to you on the mapping side. I think it was last year that T-Mobile released a very interesting product where you would punch in a street address and it would show you their own information, their own GIS (geographical information system), with an overlay of where they expected to have coverage of different types, whether it was phone coverage or EDGE or whatever service they were offering. It seems by your description, if you extended this beyond Wi-Fi, if it winds up to be viable to sell passive scanning and Wi-Max and so forth, is this something where you could see yourself providing a Google Maps-style service or overlay that says if you are in Philadelphia this pocket has terrible Wi-Fi coverage in this one block. And maybe you sell that data to Earthlink so that they know to improve it if they do not know already. It is also something where a visitor would be able to pull this up and say I'm not going to go to that park, I am going to go to this other park. Do you see these kinds of overlays being something you could offer?

Ted: Absolutely. We have been approached by the companies that do the performance testing for the wireless carries for just that reason. Those coverage maps that T-Mobile and other put out are simulated by propagation models.

Glenn: Right.

Ted: They will actually integrate GIS terrain maps with where they know cell towers are on the propagation models they have and then do some sampling to verify that, but they do not drive up and down your street and everyone knows those dead pockets in their towns where there is no cell coverage. They only learn of those from complaints. We have been asked, could we use that while we start to gather cell tower information some of that propagation information so you can have down to the ten, fifteen or twenty foot radius exactly what the cell coverage is in your area.

Glenn: That would be fascinating. An amazing competitive tool. Obviously, this is one of those perfect knowledge problems. Would all the cellular companies like you to have perfect knowledge of their coverage? Probably not. They would rather you think that their coverage is ubiquitous even if occasionally you have problems. With a metro-scale Wi-Fi perhaps they would want people to know precisely the coverage so they would not set unrealistic expectations.

Ted: It also helps them with planning. So the wireless folks, part of the discussions we are having with them is to help them with propagation and coverage. So in the client tool they roll out it will not only allow for location based applications like Friend Finder and local advertising, but it will also collect usage data for the purpose of optimizing the network.

Glenn: The QoS [quality of service] data has always been an interesting thing. Back when Boingo was introduced, one of the things they talked about, and I believe it is still in their client, is a phone-home option. Whatever hot spot you have connected to, the client collects quality service information and returns it. As far as I know Boingo has never released a public project of it but it seems like something you all could do in coordination. Your system is largely based, or almost entirely based on outdoor scanning. It seems like there could be an indoor scanning component as well that would pair with that.

Ted: Absolutely. We know that a number of the other Wi-Fi management tools like Boingo do the same thing. In order to try to help optimize the coverage in a hotel or airport they try to collect that data. We also put a bit of extra capabilities in our Loki product for example, to help with some of that indoor coverage. While you are scanning and plotting yourself on a map in your home or in your office, we are not only positioning your location based on the access points we have seen from outside, but we are also picking up new ones that are deep inside the building or up thirty or forty floors. That gets added back to the system automatically so the system continues to expand and heal itself.

Glenn: We should talk about that. That is supposedly one of the best things about the Internet, that it is self healing, although that has always been a questionable issue about whether the Internet can heal itself in the event of huge disasters like the entire eastern seaboard going dark for a blackout. Your system is designed from the very earliest days to take input back from the users in the field and integrate that with the database so you can constantly be improving the hygiene without having to drive every street every week.

Ted: Right. Exactly. From the very beginning, we have designed the system to be self-healing. The more users that come on the system, the less we have to go out and physically remap particular areas. We would then just do that on a maintenance basis. Every time someone uses Loki for driving directions or to email their location to a friend, it is updating the environment around them not only looking for access points that might have been moved, but also for new access points that have just been rolled out. Someone just went down to CompUSA and bought a new Linksys box and plugged it in, we will pick that up once someone uses Loki within that area.

Glenn: That is very interesting because it will gives them motivation to someone in a sparser area you have not driven as frequently, for users that pop up there could have a practical use for war driving to improve your system.

Ted: Exactly. One of the interesting things that have come out of the Loki deployment that we did not expect is that just under a third of all Loki users are outside the U.S. Which is where we do not have Wi-Fi coverage. Instead, we fall back to IP address location which is much less reliable. But they can use a manual update component of the product to turn the system. What that is doing, is war-driving using Loki so outside the U.S. we are picking up access points readings and their locations that users are donating to the system in effect.

Glenn: You just brought up too exceptionally interesting point. One is that everything starts in the U.S. I think because between government policy, copyright law, licensing, or whatever it is, it always seems to be a bit easier for companies to say, "I want a license for mapping information for the entire United States", "I want satellite information", what have you. So Google Maps, all the mapping services start in the U.S. and then they expand slowly to other countries as they gather information. Right now, you are a U.S. service. Do you face issues, are governments going to say we do not want you driving around with cars equipped with GPS equipment in France or Nigeria. What are the issues you are going to face as you try to expand the mapping?

Ted: There maybe some labor issues or something about how we are mapping the data. We do not know. But from a technical standpoint the standard is global. It works the same way in Guyana as it does in Australia and as it does in New Jersey. We are doing early pilots around the world right now to explore some of those other issues. It looks to operate and the results appear to be identical to the U.S. As of now, we have covered all of the cities in Australia. We are doing a pilot in New Zealand. We have done some pilots in Northern Europe. In fact, we have done one in the middle of Africa just to explore what the coverage might be in a poor country. As far we can tell there are no hurdles for us to conduct the model in the same way we have done in the States.

Glenn: You really need to partner with the One Laptop per Child program and get software bundled on it so when you have every less industrialized nation as they start using Wi-Fi in large quantities you could wind up with rich mapping data from the poorest places.

Ted: Absolutely. We are getting a lot of inquiries from international users that want to be the local arm of Skyhook out there building the data.

Glenn: I know that Estonia wants to talk to you. The man that spread Wi-Fi in Estonia would be all over that and it is a small enough country that he could probably do it himself with a few days of driving. We keep talking about Loki and I should bring that up because I think Wi-Fi Networking News readers would probably know what that is, but I have been fascinated it. You gave me a glimpse of it some months ago. You rolled it out as as beta service. It is Loki. Loki, named after the Norse god of mischief, which I hope has no connection to not finding one's directions, but it is a good name. It is Loki.com right? It is not Loki.net?

Ted: That is correct. Loki.com.

Glenn: We have talked that from the time that you first briefed me on your company--I have asked what are the applications and you have always had potential applications. And I know that there have been some partnerships with the laptop recovery company [CyberAngel] and so forth, but asked how is this going to be used and you guys have answered this question by providing what is both a rich service and a proof of concept. Let's talk about what Loki is, and what Loki offers.

Ted: OK. Loki is a free consumer toolbar. You can go to Loki.com and download it will integrate into your preferred browser whether it is Internet Explorer of Firefox. It essentially turns your laptop into a virtual GPS device. So without having to add any hardware to the system, you can now conduct all the same functions you would do with an external GPS device. And so you can plot yourself on a map, get driving directions to your next meeting or some restaurant you're going out to dinner, or just look for content on the web that is all location entagged. So some examples of that are, if you are looking for movies in your area, or you wanted the weather around you, or you were looking for the nearest hot spot or cheapest gas station. All this data is available on the web, but you have to go to different content sites and put in your address at that particular moment. With Loki, you just click a button, it figures out where you are and sends that information off to the particular website. And so what it does is it adds location to everyday browsing of the Internet and it makes the information a lot more relevant. Instead of searching Google for the entire world, you're just searching Google for an area a mile or two around your particular location.

Glenn: And the idea, you know, just like a GPS, the idea is, a GPS receiver, the idea is I'm in a typically unfamiliar place or I'm looking for an unfamiliar resource and I use this tool, and because it knows where I am in most places, it then provides all this rich information for me based on my location without me having to do, even look at a street sign nearby.

Ted: Exactly. I mean, just think about how many times you're on the road for business, or for personal travel, and you're not exactly sure where you are, other than to maybe call down to the front desk of the hotel or look on the phone. Instead of having to find that address and type it in every time you want to look up something, Loki figures out where you are and enters that into a particular website or onto a map page. And so it removes a lot of the headache of planning for a business trip or setting up a day for leisure travel. We've just found a whole number of use-cases that people are utilizing the system for from all around the country and even around the world.

Glenn: Well, you know it seems also, I know this first version, I know you're in beta. This is a toolbar for browsers, works under Windows and I've seen just looking at your site, those Linux and Macs version are on the way at some point, which is great, cause there's a lot of Macintosh Powerbook users and Mac Book users out there who are always aching for a tool like this.

Ted: Right.

Glenn: But this seems like a logical thing for a PocketPC, a mobile handheld, or a smart phone. And you know, I remember, if you remember the heady days of the Internet, when there was many different city guides, and I can't remember the name of it now, but there was a great city guide where back when the Palm had, like, 16 MB of memory. You'd download a San Francisco module, it would have restaurants and you could punch in a set of coordinates. You could say I'm at Fulton and First, and it would say, okay, take a right, walk five blocks, take a left, and you get to, you know, Yankee Dim Sum Diner, or whatever it is. So, you know, this seems like a logical extension for you guys. Is there a limiting factor in getting into a handheld format, or is it just the extent of development that you all can do, the hurdles to cross to make those programs work?

Ted: Well, it's certainly on the path for us. Not only to add different laptop or desktop platforms like the Mac and Linux, but also to start moving to these more mobile devices. And for us, it's - we prioritize in terms of the number of devices that are available today. If you were to lay them out, you'd see that there is a hundred million plus laptops out there that are Wi-Fi-enabled today. So it's a natural starting point, where you'll get the most amount of feedback and usage. But clearly the mobile aspect is where we really want to go, and the limiting factor there is really PDAs that have connectivity and Wi-Fi. And so, as you start to see Wi-Fi get into the phones and they can use the cellular data networks, it becomes a lot more interesting. And so you'll start to see us explore those kinds of platforms during the second half of the year, but we expect to see the usage spike once we start to see Wi-Fi in more of the phones.

Glenn: Right, and obviously that you're hitting two simultaneous trends, which is that people want Wi-Fi because it's more available, or they might want Wi-Fi and cell data in one device. But also the unlicensed mobile access (UMA) and similar technology that let you do phone calls over the cell network, over Wi-Fi, that's going to push Wi-Fi into more of the devices too.

Ted: Absolutely. And you're seeing that driven by the users, and so commercial customers are saying, "I have all my sales reps with these cell phones, but I don't want to have to pay for the data minutes and the voice minutes when they're in our office building. I've spent a lot of money on my data network internally, and I deploy it using Wi-Fi. I want to be able to take advantage of that. And so they're pushing the carriers pretty hard. And then we're also seeing newer wireless companies, like the MBVOs [mobile virtual network operators] who are saying, the more voice and data traffic I can push off of the cellular network, the less I have to pay the backend provider, like a Sprint or a Cingular. And so there's a lot of motivations to bring in these unlicensed tools into the wireless phones.

Glenn: And obviously there's also a market for you there in multiple ways. There's market data, but there's also partnerships. You're going to be gathering, or you have gathered more information about the Wi-Fi portfolio - what's going on with, what equipment's been deployed, where it's been deployed, what's public, what's private, than anybody else out there. No one's been doing this scale of war-driving ever. But you also have this potential, it seems like that if a Boingo or a T-Mobile or an AT&T Freedomlink or whomever, wanted to drive people to their locations, that's something that you can integrate into both Loki and into your system. If they want to let users of Loki upload location data through their Wi-Fi network, you could build authentication modules into your system that would allow you to partner. It seems like there's a lot of opportunities there that you may be exploring.

Ted: Exactly. There's a tremendous amount of opportunities for us, given the kind of data that we're collecting and just the sheer volume of it. In fact, some of the opportunities that have recently emerged is, we're seeing outside the US, there's a real problem with getting affordable map data. It tends to be controlled by monopolies, being the state governments. And so it's really hard to get street data without spending an arm and a leg. Given that we're driving every single street, we're building from the bottom up, basically map data for these major cities. And so there's opportunities for us to help data providers in other countries get map data from another source, without having to go through these central governments. And that's something that we hadn't expected when we first started out. It's just an example of some of the side business opportunities that are emerging from this.

Glenn: Well, you're developing a lot of intelligence about location, and even though it's Wi-Fi focused right now or in GPS, is sort of the outputter GPS, like the thing is the output, you've just - this is actually one of those interesting things - Companies that follow their noses for emergent properties of the data they collect often do better than companies that have a... I mean, laser beam focus is great, but there's also the "don't overlook properties right under your nose".

Ted: Right, and it's a balance, where when you're a young company like we are, between being focused and being opportunistic. You know, there's a graveyard full of companies out there that tried to chase too many opportunities, and that's certainly a risk for us as well. So we're trying to balance, you know, targeting platforms, user bases, use cases where we're seeing a lot of activity and getting a lot of feedback as well as looking into areas and turning over rocks that people hadn't thought of before, and that's just a day-to-day challenge for us in an early market like this.

Glenn: Oh, you know what, before I leave Loki too, I forgot to mention is I thought one of the interesting aspects about that ties in a little bit to the market opportunity issue is that you've built a system that's extensible from the start, so that I can go into Loki, and if I find a new website, I can create a profile for that website even with a username and password or that has to go through multiple web pages. It seems like there's a fairly sophisticated set of rules that I could build, so if I create my own resource for my own company, or if some great mapping thing comes up and you guys haven't updated it in time, I can still use Loki to punch in data into that new website.

Ted: Yes, and we've actually- we're working on some new capabilities to make that even more open. So if you were to build a website that allowed for a latitude and longitude search of your content, for example, sites like Flagr and Platial that are doing mash-ups based on user-contributed data. It's very easy for the people at Flagr or Palatial to add a channel to Loki on their own and post that on their website, and some of the things we're doing in the next version is, they could then put that as a link on their site, where someone just clicks the link and it adds it to the toolbar, so it's always there for them, and that's a very simple thing for them to do. And then in addition to that, we're also adding the ability for content sites to request a location from Loki as opposed to the user to have to proactively push a button. So if you went to the Home Depot website, it could detect that Loki was installed, and automatically on the front page, show you a map to the nearest Home Depot without ever having to ask you anything.

Glenn: Well, I'd be remiss if I didn't ask about, in this day and age, about the privacy implications of all this, is, you know, when you describe, like, you know, there's always the push versus pull argument. And when I can push something, I say, "tell me" - I'm giving you my location information - "tell me where the nearest Home Depot is", that's one set of privacy concerns. If Home Depot can pull the information from me, then obviously there's going to be some granularity in what I control. How do you set up privacy both for, you know, anonymity or lack thereof for this pull data as a consumer, or I should say "push data"as a consumer, I'm pushing out the information or the pull side, from a website that could pull the information from me.

Ted: Well, obviously privacy is a big concern for us because if we don't handle that right, we could lose all the interest we have in our technology and Loki. So, we spend a lot of time with the privacy community, actually getting them to validate our privacy policy and our approaches. And one of the things we've done with Loki, and if you go to the site you'll notice, we never ask who you are, and we never ask you to register. And that's because we want the system to be completely anonymous. As you're making requests to plot you on the map or share your location with a friend, we never know who that person is whatsoever. And we don't track that information and there's no way we can backtrack to an individual whatsoever. And that's something we've baked into the system. On the pull side of it, one of the things we've done is used conventional techniques like you've seen around cookies and passwords where when Home Depot first wants to get that location from Loki, a pop-up will show that says, "Home Depot is asking for your location. Will you allow HomeDepot.com to have this location and do you want to allow it going forward automatically?" And so you always have control it's always opt-in based. But it does give the content site a lot more control over how they tailor the content around your particular location and your needs.

Glenn: Now what about the information that gets fed back that's the readings of the scans of other Wi-Fi hot spots or access points around me. Is there an option for me to say I don't want to provide this quality of services update data because I'm vaguely paranoid maybe not so much these days, in this day and age and I don't want to give out location information about myself, even if it's just IP connected or some such. Can I disable that?

Ted: We try in the privacy policy up front to say that this is the type of data that your device is going to be sending to us, and these are the ways we're going to use it. One, we're going to use it to calculate your location, and two, we're going to use it to maintain and optimize the system. So it's really baked into how the overall system works and it's analogous to how the cell phone system works today. As you're using your phone, the cellular providers are using that as a way to improve the system and look for these dead spots. So it is part of how the system improves itself and gets better. So as of today, you're opting into that whole chain of events as opposed to just picking which pieces you want to participate in.

Glenn: Well let's ask about the anonymity part of it, though. Let's talk about that, which is, is my IP address sent as part of that? Is that stored as a permanent record that if law enforcement were to serve you with a subpoena, they could get the IP information? Or do you discard the IP information after the location details are sent?

Ted: We store those separately. We do use that because we do IP location as well for those people outside of our coverage areas, and we want to continually improve that as well. But one of the things we've done is try to segment those out so that you can't backtrack your way into looking at someone's trail or path. And so we're looking at all different ways to make sure that one, we protect privacy, but two, we want to also, for the 99% of the people who use Loki, make sure we're providing the best location. So that's a balance we continue to work on, and we're going to try to be as transparent as possible so that people know what they're getting themselves into.

Glenn: Well absolutely. It has been an issue for a long time, but it's just now blossomed into everyone's knowledge now that we know that various governments, including the US government, are actively using that information in various ways. So people think about it a lot more, and it's good to know that it's top of mind for you all as you develop, and you're disclosing what the parameters are.

Ted: One thing just to add to this we've talked a lot about Loki but the system can also work in a completely offline fashion such that you download a compressed or segmented version of the database locally, and you're never making any network calls whatsoever. And that allows you to do whatever you want without having to require a network connection but also alleviates some of these privacy concerns. And that's a model that we typically see people use for things like personal navigation, and we're looking at ways to meld that into the current Loki model so that people can almost decide their privacy concern versus accuracy need to make those choices themselves.

Glenn: Well, that's obviously one of those things when you have limited memory devices or limited power devices. If you're using a portable, this reminds me of the old Palm days. Maybe you download a several megabyte database for San Francisco or the Bay Area because that's where you're going, and you have that available to you in your handheld that has limited storage. And then, it only needs to do Wi-Fi calls maybe occasionally, or when Wi-Fi is available, or when you turn on Wi-Fi it can get the information it needs, but it doesn't have to be constantly powering the Wi-Fi unit and draining the battery power as well.

Ted: Right. And there's different techniques that we're working on in the research lab that require less battery power to do the types of scanning we're doing for location. So this is all balanced between on the one hand, delivering the end result, which is an accurate location reading, and balancing that against a user's privacy concerns, against things like battery and storage and network connectivity, and finding the right balance based on the particular needs of the user at that time and the capabilities of the device and network.

Glenn: I should ask one other privacy question which is a little obscure, but it's that you're scanning Wi-Fi signals passively to see what's out there and use that for location information. I'm assuming that there's no way as an individual that I could opt out of having my access point included in your database because you're not exposing my access point as one of the many millions that you're scanning from. That information isn't being produced, I assume, exposed in some fashion to users.

Ted: Right. Our users never see the underlying data whatsoever. So while we're gathering information about your personal access point, no one's ever able to use that in any way to exploit your system. In addition to that, we do try and be very open about it. If somebody has an extreme concern about their access point being included, we're more than willing to take that out of the system to alleviate that concern. But folks do need to be concerned that when they use a technology like Wi-Fi, they are using a device that operates in the public spectrum. And so there are certain obligations and certain trade offs that you make when you do that. Wi-Fi is very cheap and very effective, but you're also sharing the airwaves with other people. So you don't necessarily have rights over that signal that you're shooting out into the air.

Glenn: Well, let me ask you one final question, since we're talking about signals shooting out, and this is a future orientated question. How does 802.11n and MIMO, how does that affect your ability to position? Does that make your job harder or easier, with longer signals, with stronger power over greater distances?

Ted: It's a double-edged sword. One, it helps with coverage, so with more.11N devices, you're assured more coverage and less gaps in maybe those suburban areas. On the downside, the cells get wider because the power is stronger, and actually the trickiest thing that we have in doing positioning is that the range of access points is actually a lot further than people think. Most people talk about the hundred yards that Wi-Fi will go, and they'll always talk about how they can't get a signal in their kitchen or basement, but that's primarily for doing data connectivity and checking your email and browsing the web. But the signal beacon itself, which is all that we use, can go very far, and in fact, we have many access points that are one, two kilometers away. And those are the biggest challenges for our positioning, and where we spend a lot of time on the algorithm side, is knowing which access points are N, and have a very wide footprint, and those that have a much tighter footprint, which we prefer.

Glenn: That's very interesting. That means you could have seamless coverage, but it might be even harder to differentiate where things are.

Ted: It just means that we have to know which ones to trust in which situations. And so we do that a fair amount in adding quality attributes to each access point that says, if you have three small radius access points, use those, because you'll get the best possible reading. But if you only have two, pick up the wider one, but weight it a little less. So those are the techniques that we've developed and built patents around, because that's what produces the best end result.

Glenn: Well, that is a very interesting glimpse into the future then, too. Great. Well, thank you so much for your time today. I've been talking with Ted Morgan, the head of Skyhook wireless and the makers of the Loki toolbar for providing GPS-like information using Wi-Fi. Huge database they maintain. Thanks, Ted.

Ted: Thanks, Glenn, good talking to you.

Glenn: Well, that was podcast number one from Wi-Fi Networking News. This is Glenn Fleischman, the Editor of that website. You can find us at wifinetnews.com. This podcast is copyright 2006 by Glenn Fleishman and it's a product of Wi-Fi Networking News, a site in which I try, every day, to provide some insight into what's going on in the industry around us. Wifinetnews.com one-stop shopping. A lot more podcasts in the days and weeks to come, including podcast #2 with Frank Hanzlik, the director of the Wi-Fi Alliance, coming at you soon. Thanks for listening.