Tuesday Jan 02, 2007

Drilling Down in the Solaris Registrations Map

Recently Jim Grisanzio referenced our Solaris Registrations Map to illustrate the volume of OpenSolaris activity in Japan. Since I'd love to see more people using our map in this way, I thought I'd talk about it a bit.

For this kind of use, one handy feature is the ability to reference a particular map view directly. This can be done by copying the "Link To: ... This View" URL when you have the map zoomed and positioned as desired. For example, Jim could have referenced this URL when talking about Solaris-related activity in Japan. This allows everyone to look at approximately the same map view and statistics. (Note: the URL controls the map type, zoom level, and center coordinates; it obviously cannot control clients' screen resolution, which is what determines how much of the map around the center is shown and leads me to use the "approximately" qualifier.)

On my screen, this view currently shows:

Registrations In Visible Area
Solaris 10 / sparc:1715
Solaris 10 / x86:6387
OpenSolaris / sparc:17
OpenSolaris / x86:290
Total:8409

How do we interpret these numbers? Well, one thing they do not mean is that there are just 8409 Solaris users in Japan. As the FAQ notes, this map only shows data for "Solaris 10 and Open Solaris instances that activated Sun Connection to receive automatic software updates." As with any product, only a subset of total users will go through a registration/activation process.

I'm not aware of a good way to estimate what percentage of total users will have registered. So I can't infer the total number of Solaris and OpenSolaris users in Japan from this map view. On the other hand, it seems likely that whatever percentage of users choose to register/activate in one region would roughly equal the percentage of users who do so in other regions. If that's true, we should be able to use this map to compare the relative size of Solaris and OpenSolaris users in different geographic regions. For example, a fully zoomed-out view of the map currently shows a total of 83268 activated instances. So comparing our Japan total (8409 instances) to this number, we could estimate that around 10% of worldwide Solaris and OpenSolaris users are in Japan. That's interesting (and makes me think that perhaps we should update the map to show such percentages automatically).

Hopefully this gives you some ideas on how to dig for interesting views and stats in the map. As we've seen, the statistics it shows are not good indicators of the number of users or installations in absolute terms, but may be useful in estimating the relative populations for different geographic areas. We plan to investigate adding new data sets (such as Jim suggests) which may provide more absolute population info. Stay tuned. And let me know if you have ideas for a data set you think should be included.

Wednesday Dec 20, 2006

The Story Behind The Solaris Registrations Map

I'm one of the people who created the Solaris Registrations Map which has caught Jonathan's eye. Developing the map has been a lot of fun, and I thought I'd share some highlights here.

First off, credit where credit is due... The original idea for the map came from Steve Wilson, Vice President of the "SysNet" group at Sun. Among other things, this group runs the Sun Update Connection and the related Solaris registration service from which the map's data comes. Steve had only been on the job a few days when at small group outing he thought aloud: wouldn't it be cool if we made a Google Maps mashup of all of these registrations we're getting? As another of the map's developers would later say, to some of us Steve's question was the equivalent of chumming the water for sharks. We knew a great idea when we heard one, and we just had to make it happen.

So over the next few days and weeks, we started playing with visualizing our registration data using the Google Maps API. A few challenges quickly became apparent. Chief among these were performance and privacy concerns.

On the performance front, we quickly realized that the JavaScript-based markers often used for such mashups just would not work for a large data set such as this one (80,000 registrations and counting). So for our high-level views, we use the Java Image I/O API to create our own matrix of tile images, and then use Google Maps API calls to have those overlaid atop the normal map tiles. For example, where a fully zoomed-out view of the earth would cripple one's browser with the work of placing 80,000 individual "balloons" on a map using JavaScript-based markers, the custom tile approach means the browser just has to retrieve and position at most a handful images. Much faster and kinder to your browser.

Next, the privacy concerns... We had originally been assuming this map would only be available to a restricted audience on Sun's internal network. But soon after we started showing off initial versions, certain people were asking why we couldn't expose it on the Internet. Again, we knew a great idea when we heard one and had to make it happen. After all, Sun has jumped into the spirit of transparency and the Participation Age like no other. Making our map public could be one more step in that path, but it had to be done with care: we absolutely could not violate the privacy expectations of our users. As cool as some people might think it would be to find a dot on their exact house, others would understandably be quite upset.

And so, after consulting with Sun's privacy specialists we settled on an approach of locating registrations by nothing more specific than a zip or postal code. In other words, your dot won't show up directly on your house (unless, perhaps, you happen to live at a central post office). Instead, information from you and any of your Solaris-using nieghbors will be aggregated together and put into one marker at the center of your postal code area. As our team's Director, Eric Peterson, put it: even "Ted Nugent would be comfortable knowing that the T1000/Nevada registered to his 'Tedquarters' in Jackson, MI is untracable beyond 49202."

I think there are some interesting details deeper down in our technical approach, but I'll cover those in a future post. In the meantime, it looks like Steve has posted his own account of the map's story.

About

woodjr

Search

Archives
« April 2014
SunMonTueWedThuFriSat
  
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
   
       
Today