USGS opens gateway to Earth

Before the Geological Survey teamed with Microsoft Corp. to set up the TerraServer Web
site, USGS had never tried to put so much data online over the Internet.


The earth science agency typically distributes its aerial image data on CD-recordable
disks or 8-mm tapes, or stages it over the network using the File Transfer Protocol.


“The people at Microsoft wanted data the public would be interested in,” said
Hedy Rossmeissl, senior program adviser at the Geological Survey.


“They originally thought about Dow Jones data and genealogy data sets. Then they
stumbled on these aerial images we had,” she said.


Under a three-year cooperative R&D agreement, a dozen USGS employees have worked
for a year on the Microsoft TerraServer project, processing 3 terabytes of digital
orthophoto quadrangle images.


“This is a lot of data, and being able to section it and sample it and serve it
back and forth across the Internet is a huge challenge,” Rossmeissl said.


The earth’s curvature presents special problems.


“When you’re trying to stitch together geospatial images, you’re always
running into little snags about how they fit,” she said.


Since June 24, the processed images have been posted publicly on an Internet-attached
64-bit AlphaServer 8400 supplied by Compaq Computer Corp.


The public server has eight 440-MHz Alpha processors. It runs Microsoft Windows NT 4.0
Enterprise Edition and the third beta release of Microsoft SQL Server 7.0 Enterprise
Edition.


“With TerraServer, we learned a huge amount about very large databases and are
feeding that back into our development process,” said Jim Gray, senior researcher at
Microsoft’s Bay Area Research Center in San Francisco.


Microsoft officials said the 41'2-ton server has 10G of memory and eight
cabinets’ worth of Digital StorageWorks ESA 10000 RAID Level 5 storage.


The server’s four NT physical volumes each hold up to 600G.


A server-attached StorageTek 9714 DLT tape library from Storage Technology Corp. of
Louisville, Colo., stores 3.5G of uncompressed image data, which it can transfer to the
disk subsystem at rates up to 108G per hour. The library has six StorageTek 7000 digital
linear tape drives and 100 cartridges.


A new version of Legato NetWorker software from Legato Systems Inc. of Palo Alto,
Calif., backs up the TerraServer database to a 12-terabyte StorageTek TimberWolf 9710 tape
library.


NetWorker runs with the Legato BusinessSuite Module for Microsoft SQL Server, a
database-aware backup and recovery application that performs online backup of the SQL
Server database, Legato officials said.


The TerraServer, located at Microsoft’s Internet data center in Canyon Park,
Wash., links through a 100-Mbps Ethernet switch to a half-dozen four-way Pentium Pro
servers running NT and Microsoft Internet Information Server 3.0.


A pair of 7500 series routers from Cisco Systems Inc. of San Jose, Calif., manage the
heavy traffic from two incoming OC-12 circuits and 11 T3 Internet circuits.


The TerraServer is the first in a series of Geological Survey projects to build a
gateway to the Earth, said Barbara Ryan, associate director of the 119-year-old agency.


The digital orthophoto quadrangles are aerial photographs, “but a little
different,” she said. “Each photograph has been rectified, and any distortion in
what you would normally see in an aerial photograph has been removed, so what you’re
looking at is like a photograph of a map.”


Digital orthophoto quadrangles, she said, make a framework for plotting elevation data
and other information about water and transportation resources.


Federal, state and county agencies use digital orthophoto quadrangles as base maps for
analyzing wetlands, soils, land parcels, farm field boundaries and forest inventories.


The digital orthophoto quadrangle images are part of the National Spatial Data
Infrastructure framework, a project of the Federal Geographic Data Committee that sets
standards for collecting and registering geospatial data.


The millions of black-and-white digital orthophotos that the public can download free
from http://www.terraserver.microsoft.com are low-resolution, 10K Joint Photographic
Experts Group-compressed tiles.


The full-resolution, 55M digital orthophotos suitable for geographic information
systems applications are available for online purchase from a separate Web server running
Microsoft Site Server 3.0 Commerce Edition at the Geological Survey’s Eros Data
Center in Sioux Falls, S.D.


On the same day the TerraServer got 4 million hits, the USGS server in Sioux Falls
registered about 100 hits, Rossmeissl said.

inside gcn

  • data science (chombosan/Shutterstock.com)

    4 steps to excellence in data analysis

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above