MapPoint Location Server debuts (Part 1)


Well-Known Member
Spent last evening at a reception Microsoft held in San Francisco to celebrate the debut of its MapPoint Location Server.While everyone else was drinking and having fun, I was sitting at a table in the back of this Italian place where the Stones are supposed to have partied, talking to folks from Sprint, Cubistix (who, oddly, don't list a URL on their business cards), and Microsoft. Mick picked a good place--the food was excellent. Here's my report:

This is the first of several pieces I am going to write about Microsoft's new MapPoint Location Server. This is a server that sits at the customer's location (typically) and powers applications that take location information provided by wireless carriers (where is this phone?) and merge it with mapping information (show me where the phone is).  The mapping data comes from Microsoft's MapPoint Web Service.
This allows--to use a Cubistix client as an example--a bakery to reroute its delivery trucks in real-time when an urgent order comes in. A major news organization is using this technology to keep track of its satellite trucks, crews, and correspondents in order to dispatch the closest resources when news happens. This has been possible--sorta--in the past if you were willing to invest enough money in the project. What Microsoft and its partners are doing is making this much more widely available, at reasonable prices, and using tools developers already understand.
I agree with people who described this as a “watershed event“ in making location-enabled applications available to American (and Canadian) businesses at prices they can afford. If your business involves sending people to customer locations, whether on service calls, deliveries, or “when news happens“ this is technology you can use. Maybe not immediately, as we are still a bit on the bleeding edge, but now is the time to get this on your radar and plan for 2005 and beyond.
At the announcement, Microsoft was joined by Sprint and Bell Mobility (Canada), who are the initial carriers providing location information. That's what I'd like to delve into a bit in this first installment. The following is based on a 45-minute discussion with Hugh Fletcher, product development manager, in Sprint's business solutions group. (These notes were taken in a darkened room on a scrap of paper, so if Hugh tells me I got some of this totally wrong I'll happily make the corrections).
(His email address is hugh.h.fletcher at Sorry for not embedding the link, but I don't want his email harvested from this post.)
How does your cell phone know where you are?
For a couple of years, Sprint has been selling wireless handsets with a GPS chip inside. This provides what I believe to be rudimentary GPS location fixes from the handset to the Sprint network. The GPS inside your Sprint handset does not work as well as a handheld GPS. It is much slower to update and doesn't provide a highly accurate fix. Its real purpose is supposed to be helping E911 dispatching, though a lot of this sounds like voodoo to me.
The Sprint network combines the handset GPS information with where on its network you are calling from, down to a particular cellular site. Together these technologies will track you to within about a mile radius. By the end of the year, it should improve to a 5-50 meter radius. By next summer--when a new Qualcomm chipset is available--your phone should be providing as good a location fix as a standalone GPS device. (The issue right now is power consumption, which the new chipset will lower to acceptable limits).
I start with this because it's important to recognize what these applications can and cannot do. The current location information is too coarse for applications that need to manage a number of resources in a fairly close space. It will not support a life-and-death app like ambulance dispatching (although there are other systems that work fine--we're talking mass availability of location data for corporate applications here, not what you can do if you have lots of money). What's currently available is fine for a company with perhaps two dozen bread trucks spread out over a 1,000-square mile area.
When the new Qualcomm chips become available, the network will receive a precision fix from the GPS-enabled wireless handset and it will become possible to manage a larger number of resources (vehicles, people, etc.) in a much tighter area.
This, of course, presumes you have a GPS fix available. Getting one requires a view of the sky and between 2 and 4 satellites in sight. If you are standing in an open field this isn't a problem, but put anything between you and Heaven and your GPS fix might go away. Go inside a typical office building and count on it, though your cellular coverage may be lost as well. When that happens who cares where you are--they won't be able to contact you anyway. Sitting in a vehicle with your phone clipped to your belt or in a briefcase can also lose your GPS fix.
When a GPS fix isn't available (the phones save the last known location and will send it), the wireless network uses its own means (triangulation of the radio signal from the handset, I believe) to work out the best fix it can. This is where the one-mile radius comes into play, although by year-end we should be down to 150 feet or less.
This also brings up the limitations of the wireless networks themselves. This location technology only works if you are on the network. Coverage may--emphasis may--be solid in urban areas, but out where I live it can be pretty spotty. If all you want to know is where your TV satellite truck was the last time it was in the cellular network, maybe 10 minutes ago, this won't be a problem. But, again, it's not good enough for life-and-death applications.
I started with the coverage limitations for the wireless network and the limitations of the ability of current technology to generate a precision location fix because these determine what applications are possible. Basically, if knowing the location of a handset and whoever/whatever is attached to it would help you create a useful application, it's now possible or soon will be.
You can't follow the device everywhere (takes a satellite uplink for that) or know precisely where it is (takes a real GPS out in the open for that) but even with these limitations, there is an awful lot you can do.
In my next post (in this series) I'll talk about how the Microsoft solution goes together. If you know of competitors I should be looking at, drop me a line.