to get from their exploration ships and have processed at several supercomputing centers in the U.S.," she adds.
Amoco, the originator of the project, searches for oil by mapping the earth with complex seismic computer programs that generate files of gigabyte to terabyte size. This data often comes from remote sites ranging from fields in Siberia to ships at sea. Amoco must move this data from the remote sites to central data-processing sites in the U.S., including the Minnesota Supercomputer Center in Minneapolis.
Traditionally, data communications between ships at sea and processing centers have used what can only be described as the world's most expensive sneakernet. Exploration ships collect information from promising locations, transfer the data to reels of magnetic tape, and fly them via helicopter to onshore facilities for processing (
see the figure
). It can take days for the dat
a to travel over leased lines to other supercomputing centers for analysis. Because the distribution process takes so much time, exploration ships usually don't stay in the exploration area awaiting the results of the analysis and instructions on where to proceed next.
Amoco's network currently uses time-division multiplexing (TDM) technology with a mesh of T1 links that serve 400 domestic and nearly 50 international nodes. Four networks are mapped over the physical TDM network for Systems Network Architecture (SNA) traffic, voice traffic, X.25 links, and routed data traffic. Amoco's LAN environment supports TCP/IP, DECnet, and Fiber Distributed Data Interface (FDDI) for LAN backbones.
When the data arrives at the supercomputer center, Amoco's researchers in Tulsa, Houston, and suburban Chicago use the network to gain remote access to seismic modeling and simulation applications. These applications produce large volumes of output, for which ATM may well be suited.
ARIES is testing terrestrial conn
ections that link together all the project's computing facilities at 45 Mbps using an ATM network supplied by a number of service providers. Added to the land-based network is the ship-to-shore link using the Comsat ATM link accelerator to collect and distribute data in real time. "ARIES participants are having to reduce the cost of oil exploration--one of their major cost centers--by a factor of 2 to 4 times, depending on how well they can make the ATM network run.
"Better communications may make it possible to reduce our cycle times dramatically for oil exploration," says David R. Beering, who is codirector of the ARIES project. "Instead of sending a crew to the field to make expensive test drills, we want to do more [work] on the computer. Doing so could increase our probability of finding oil, because we will have better information about where to drill."
In addition to reducing exploration time and the associated costs, Amoco will use the project to test ATM as a way to enhance its corporate tel
ecommunications network. The $25 billion energy company runs operations throughout the world over an enterprise-level network that typically doubles in size every 10 months. Ethernet, token ring, Fast Ethernet, and FDDI simply aren't up to the task of handling this mushrooming data load. As a result, Amoco's telecommunications experts are evaluating ATM's ability to provide bigger data pipelines to help the company more efficiently fill its oil pipelines.
illustration_link (28 Kbytes)

BENEFIT:
ATM in the petroleum industry's ARIES project has redduced costs, cut exploration cycle times, and
helped one oil company expand its telecommunications networks.