Geo engineering data turns into gold
S. S. K. Banerjee
[email protected]
There is no such thing as disposing of infrastructure. Go to Rome and you can drive on the Via Apia, a 2000-year-old pavement. Infrastructure is cumulative; we don’t remove utilities, harbours, buried sewers, etc. We add more, and more, and more. When infr A structure is physically destroyed, the scale of the demolition job is of the same magnitude as the initial construction job (consider the dam-removal programs being carried out in some parts of the world).
br>
Creating the mother lode
As we keep adding infrastructure, we also accumulate data relating to the entire existing infrastructure. The investment associated with the accumulated information is enormous. It has represented, typically, over 80 percent of all GIS-related project cos ts. Even if Moore’s law holds true, the doubling of computing power every 18 months pales when compared with the exploding volumes of the data we collect every year. To some extent, it is just that increasing computing power makes our data-collection poss ible, from mobile video capture, to high-resolution scanners, to satellite imagery. Space Imaging’s one-meter-resolution satellite, now due next June, for example, will generate 40 megabytes of data per second, every second of every day.
Paralleling the growth of newly-captured data is digital information, previously private by nature, that is now becoming open and accessible. Many agencies and organizations have volumes of data on servers to which they had no practical way to provide any one access. Under those circumstances, there was little debate about who should have access to that data, simply because there was no technology to make it accessible.
The Internet is changing that. To publish this data on the Web does not require much work. The data exists already in digital form, maintained properly. Today, servers containing massive amounts of data can be reached by any authorized browser.
Mining the data
Given the theoretical availability of this information on the Internet, how practical is that form of access? Granted, data publishers provide CDs with hundreds of megabytes of data to you, and there is no way for the Internet to replicate the way CDs are distributed. And on the horizon are DVDs. But over the Internet, the question is usually stated in terms of raw bandwidth. With the bandwidth I have from my Bentley connection today, I can realistically download files as large as 20 or 40 MB (and I can continue to do my work with the download pro gressing in the background), but that does not allow me to replace the FedEx envelope containing a six-pack CD.
There will always be greater demands for bandwidth, and bandwidth will continue to improve. There are other factors that can change the way we work with remote data, and ease the bandwidth requirements:
- Pertinence. Most users are interested in a small data set (either spatially located-“show me all of the buried pipes around that area,”-or pertinent to a specific theme-“show me all of the operating oil wells in this Province”).
