AAPG Site Search | Home > EXPLORER > Archives > June 2002 > PC challenges UNIX
Classifieds
Advertising
AAPG Annual Meeting 2002

GO TO:

UNIX -- The Challenger is Challenged

PCs' Muscle Makes It a Contender

By KATHY SHIRLEY
EXPLORER Correspondent

Several Factors Can Influence System Decisions

There are plenty of factors in the Linux vs. Windows debate.

Both have advantages, according to Murray Roth, vice president of research and development for Landmark Graphics.

  • Linux's biggest advantage: It's the same technology that's been used on Unix and can actually work in a hybrid environment, making the transition from Unix to Linux smoother.
  • The biggest advantage of Windows is its familiarity.

"People have been using Windows on its business desktop for years, so the workflows are the same," Roth said. "Both these systems offer the capability to bring together the business and technical aspects of a company."

Of course, despite the increasing advantages of PC technology, there are corporate barriers -- and among the most important is the cost associated with this type of massive shift to a new system. Most companies are still heavily focused on Unix technology because the firms have a huge investment in those systems, according to Jim Sledz, director of exploration information management strategy with Conoco.

"With the investment in Unix we have to look at this closely," he said. "When the end users start touting moving to a PC-based system you have to challenge them. What are you getting that you can't do on Unix today to justify these costs? Right now there's not that much. PC systems are faster, but I can get a faster Unix machine for a lot less than what I have to invest in supporting a whole new PC system."

Also, the issue of limited memory on PC systems is critical for oil companies.

"Data volumes are constantly increasing, and everybody wants to see all that data on their workstation screen at the same time," he said. "To get these bigger volumes and bigger cubes you have to use Unix.

Bottom line for him: "There's going to have to be that leap in memory capability for PC systems before it can make any real inroads."

Often it's replacement cycles that drive oil companies' needs, he added.

"If I just bought a new system, then I'm not going to be interested in changing," he said. "But if my company is at the point of changing out the corporate system, then I'm going to be interested in any new systems that can improve my capabilities.

"These are business-based decisions within large oil companies. You can't suddenly just decide to spend $5 million to convert to a PC-based system," he continued. "I have to put a business case forward and compare that with other issues within the company, so I better have a compelling argument.

"You have to take a hard look at what you're getting for that investment, and what is the rate of return."

-- KATHY SHIRLEY
 

As dawn broke on the 1990s, geoscientists at large oil companies around the globe were heralding desktop Unix computer systems, the latest breakthrough in computer technology, as a creation that would soon send massive centralized workstations the way of the dinosaurs.

In fact, that shift materialized.

But today, just a mere 12 years later, Unix technology is being challenged by that other unassuming machine everyone has on their desktop both at the office and at home -- the personal computer.

Welcome to the always-evolving world of computers.

"The magic of Unix was that for the first time everybody could have a machine on their desktop -- there was no more fighting over the centralized workstation," said Murray Roth, vice president of research and development for Landmark Graphics. "Suddenly geophysicists had interpretation and other capabilities right on their desk."

But then the geologists and reservoir engineers and others wanted the same capabilities, so additional applications were added.

Unix technology was a great enabler for the expansion of software technology that gave depth and breadth to the exploration and production process.

"In the past, PCs were never a contender in this arena because they were slower, less capable central processing units than Unix," Roth said, "but PC advancements have far exceeded the growth of Unix, and today are actually at a crossover point.

ÓAdvanced personal computer hardware, along with dramatically improved graphics capabilities -- thanks to gaming technologies -- have made PCs a viable alternative for all geoscience applications," he said.

"There are fewer and fewer barriers to the wide proliferation of PCs as an option."

Gaining Momentum

Bret Fossum, a senior geological advisor with Conoco in Houston and the chairman of AAPG's Computer Applications and Internet Committee, said the transition to PCs has begun within oil companies because of the efficiencies that the user community gains with PCs.

"It's not a total solution yet, but it's gaining momentum," he said of new PC technology. "A PC platform provides general efficiency. You don't have to deal with two computer systems."

Plus, he added, PCs offer the advantage of portability.

"I have a laptop I consider a workstation," Fossum said. "I can go on the road and make a workstation-type presentation. You can't do that with Unix."

Also, PCs compare favorably with Unix systems in terms of speed.

"When companies are looking to staff projects and want those teams to be as efficient as possible, making computers faster is important," he said.

And increasingly complex software programs demand bigger, faster machines.

"For example, 3-D visualization is a common application in the geophysical and geological world today," Fossum said. "A typical Unix-based visualization system runs $1 million with three projectors, multiple processors and other hardware. You can get a fairly decent PC-based visualization system for under $30,000.

"There is the potential for some serious cost savings, depending on what you want to do with the system."

Let's Talk Money

Today some geoscience software companies are building their tools exclusively on a PC platform.

When the founders of Petrel established the company in 1996 they decided to build the company on the Windows platform, according to Erik Wulff-Pedersen, sales manager for the firm. Software development turned out to be much more efficient on Windows, and the rapid development of PC-based processors and graphic cards convinced the firm that PC/Windows was the future for 3-D modeling.

"The PC solution is so much better from a user standpoint, because everybody is familiar with Windows in general," Wulff-Pedersen said. "That means less training for new hires and a smoother workflow in all aspects of business.

"Plus," he added, "PC processors and graphics cards are now faster than Unix systems."

In addition to those advantages PC-based systems bring to the end user, programmers also claim they can program up to 10 times faster on PCs compared to Unix.

"That's a big advantage," he added, "because it allows us to develop and implement our tools much faster."

But what is the biggest advantage for PC-based systems?

Price.

"We've conducted a price performance analysis on a wide range of applications and across the board we are seeing a 10-fold baseline price performance difference," Roth said. "We now have PC systems that are two to three times faster and often five times less expensive than a fairly comparable Unix machine -- that even surprised us.

"This is going to be a staggering value proposition to information technology departments and end users," he added.

"Today we can cut IT costs by a factor of 10 and can run the exact same applications you have come to love over the years on the Unix systems."

Some Potential Problems?

While PCs have gained tremendous ground on the traditional Unix systems, there are still problems that must be overcome before the industry sees wide proliferation of PCs for geoscience applications.

First, memory is an issue. Today PCs are limited to an effective 32-bit operating systems compared to 64-bits with Unix.

"The 64-bit operating system gives you the capability for a great deal more memory and large seismic volumes, for instance, require a lot of memory," Fossum said.

Roth agreed.

"PCs today are confined to four gigabytes of memory because of the 32- bit operating system architecture," he said. "When PC hardware capabilities reach 64-bit -- and that will be soon -- there will be very little reason to be on Unix."

But another barrier to PC proliferation within oil companies is that most firms store their data in the Unix environment.

"In the absence of that link to the corporate database we build our own databases on our PCs -- but that's not very efficient," Fossum said. "We need to be connected to the corporate database so the work can flow back and forth."

"Plus, it's important to retain all the corporate knowledge base," he added, "so a link must be developed to allow all work to be maintained in the corporate database."

Wulff-Pedersen agreed that there has been some reluctance to go to the PC world in oil companies because PCs don't have the database capabilities that have been generated in Unix systems.

"I don't think we will see any databases for PCs in the near future," he said, "but as long as solutions in the form of links can be developed, then that problem can be overcome."

Such a link will become available this year: Open Spirit is a link developed specifically to connect Unix and PC systems.

"With this link an end user can retrieve data from the corporate database, conduct their work and then read the data, along with any changes made to the data, back into the corporate database," Wulff-Pedersen said. "In this way everybody is working with the same dataset."

Open Spirit was developed within Shell and other companies, but today it has been spun off into a stand-alone firm.

"Open Spirit is creating a layer to move data smoothly between the corporate database and applications -- it doesn't matter if it is a PC or Unix application," Wulff-Pedersen said. "This is a huge benefit, since it eliminates the need to create another PC-based database."

To justify the costs associated with a massive changeover to PC systems there has to be fundamental changes that move technology forward in terms of applications and functionalities.

For example, a pilot program to utilize a cluster of dual processing PC class computers is being developed at Conoco in Houston, to be used in parallel computation for 3-D visualization projects.

Over time more fundamental changes will occur, according to Jim Sledz, Conoco's director of exploration information management strategy.

"If you ask five oil companies what their timeframe for this type of change might be you would get a range from immediately to three years out," he said.

Roth agreed that, while he is seeing real excitement from the information technology community for PC-based systems, the change won't come fast. "Unix will still be here 10 years from now," he said, "but I do think a substantial number of users will be transitioning to PC-based systems as early as the end of this year and going into next year."

Wulff-Pedersen said that "we are developing tools today for the Nintendo generation -- and that's a good thing.

"The workflow is becoming more like an interactive game than a struggle to get bits and pieces of information into a model," he said, "and that enhances creativity and new ideas."