Issue link: https://read.dmtmag.com/i/53167
In the imaging industry, the time has come to rethink how to handle "big data" (particularly imagery, which is the "biggest" data of them all). The old paradigm is to ship imagery around and keep copies locally. This works when data volumes are small, but it clogs up communication pipes and creates massive—and expensive—needs for storage, power, cooling and data- center real estate. Accordingly, the industry is rapidly evolving the imagery model to move from replication to centralized hosting, with local caching for speed, to meet customer needs and deliver cost savings. First, with such large datasets in the geospatial world, it's important to move the computing to the data, so end users aren't forced to transport—or manage— the raw data. Second, mobile computing fits naturally in a model where the heavy lifting is done on the server side (i.e., in the cloud, along with the data) instead of on a mobile device that generally runs a lightweight "app." Because these two trends reinforce each other to drive the creation of a rich set of cloud geospatial- computing capabilities, it's natural to see even high-end geospatial applications moving to take advantage of this. With all the attendant advantages in simplifying the IT infrastructure—such as deploying software updates A satellite image of Japan shows damage at the Dai Ichi Power Plant after the devastating earthquake and tsunami in early March 2011. centrally to the cloud, instead of having to deploy too many workstations—this reduces the cost of managing software updates as well as the time it takes to imple- ment them. As imagery shifts from local copies to cen- tralized services, having a rich collection of metadata to enable search and discovery becomes vastly more impor- tant, as it becomes critical to find the correct image(s) to not clog network bandwidth by shipping vast quantities of irrelevant images simply to find the one(s) that matter. So efforts to extend the richness of image tags take on much greater importance in this new world. Today, satellite-imaging companies tag imagery based on time and geometry: when was the image taken, where was it taken and a few content param- eters (e.g., how much cloud cover). As the industry moves toward a hosted model, it will need to tag images much more based on what they contain. Many 2011 events in the Middle East, now known as the Arab Spring, were captured by satellites. An image from February 2011 shows a massive crowd gathering in Tahrir Square. 28 G EO W ORLD /JANUAR Y 2O12 A True Digital Globe Looking a little farther out, a second major trend should be the creation of a trusted, global, visually based reference platform. The Earth is continuously changing, and that change is accelerating, meaning a tremendous increase in data. So how can customers trust that the information they're accessing is current, complete and accurate? Imagery/LIDAR Special Issue Satellite Imagery