I got a real “two-fer” when I spent some time with Deke Young, the Director of Business Systems at GeoEye. I got the inside scoop on the back end of its new GeoFuse applications and I learned more about the cloud and its use.
First, GeoFuse. It’s the next generation of GeoEye’s old ImageSearch, the tool used to find, evaluate and then purchase imagery. That first generation app was built by Canadian company Compusult. With the imagery from GeoEye-1 coming online, last year Young and his team decided it was time to be sure the app truly made the company’s products “timely, accurate and accessible.” To do that it looked to key technology providers for the backend: Microsoft for data storage in SQLServer, ESRI for GIS functionality via ArcGIS Server and Google for its familiar interfaces. Put those together and you have GeoFuse.
Once I had that in mind Young switched gears and started talking about catalog retailer J Crew. I know, but stay with me… J Crew has a warehouse with all sorts of products. Instead of sending everyone a huge catalog, it offers men’s, ladies, summer, kids, etc. The idea is to make it easier for different buyers to find just what they need. Take that logic back to GeoEye and the list of possible ways to search for imagery (Google Maps, Google Earth, ArcGIS) makes perfect sense. In each case the same queries are sent to the same ArcGIS Server implementation (three nodes actually) hosted at GeoEye. The only difference is the nature of the client. In ArcGIS you query with ArcGIS type tools to say, sort imagery by date. In Google Earth you can use the built-in slider bar to have them pop up as you move through time. These apps, I found on quick look, are a “best practices” way to think about serving different users geodata and functionality.
Now, a cloud story. GeoEye had planned to share an image of President Obama’s inauguration from GeoEye-1 back in January. Once the image and its availability was shared on CNN, demand through the GeoEye pipe to the Internet grew, threatening to block it (think of a plugged up pipe!). Many, many people wanted to download that 3Mb image. But GeoEye was ready. A quick phone call and a credit card moved the image to Amazon’s S3 (storage services), and a quick change to the website pointed requesters to the new URL. Catastrophe was averted by the cloud! Now, all of GeoEye’s sample images live up on Amazon. In time, some processing might be up there, too, says Young.