The future of the enterprise and grids

The future of the enterprise and grids

Summary: Remember all the talk about outsourcing and offshoring, building virtual corporations in the same way Hollywood assembles disparate groups to make a movie? That’s become passé now, or at least more of a background to the larger issue of globalization.

SHARE:
TOPICS: Outsourcing
1

Remember all the talk about outsourcing and offshoring, building virtual corporations in the same way Hollywood assembles disparate groups to make a movie? That’s become passé now, or at least more of a background to the larger issue of globalization.  Enterprises are shifting their focus to have footprints around the world under one umbrella of accountability, according to Bill Janeway, vice chairman of Warburg Pincus, who hosted a discussion at the Future in Review conference with AMD CTO Fred Weber and Dick Lampman, senior vice president of research at Hewlett-Packard. The Internet is enabling easier communications across geographies and time zones, but  that doesn’t mean that face-to-face meetings will fade away. "IT makes it possible to work globally, but in no way makes it easy," said Weber. "Telepresence is great, but the people you go out to dinner with will forge a deeper relationship." Lampman chimed in that the Internet provides a "set of power tools to build on top of human relationships and trust."

freddick.jpg

AMD CTO Fred Weber (left) and HP SVP Research Dick Lampman

An IT global community, both internally and extending to partners and outsourcers, creates more organizational complexity. "The complication of global business is as big as the opportunities it creates," Weber said. One of the dangers is bringing every decision to a central authority, which slows down decision making. Technology can help speed decision flow, but not if the organizational structure is bottlenecked. 
 
In the area of new models for transacting, HP is investigating auction models and dynamic pricing, both for selling to customers and filling its own supply chain, Lampman said. 
 

John Thompson, CEO, Marketing Sciences, hosted a panel of experts on the future of grid computing. Panelists included Wolfgang Gentzsch, managing director, MCNC; Wayne Clark, architect, Intelligent Network Services, Cisco; Larry Smarr, Harry E. Gruber Professor, Department of Computer Science and Engineering, UCSD, and director, California Institute for Telecommunications and Information Technology. Panelists highlighted the merging of grid protocols and Web services and alliances among various standards bodies and trade groups, such as the Global Grid Forum and the Globus Alliance.   

larrysmarr.jpgLarry Smarr noted the problem with grids and the shared Internet infrastructure. "We all live in little data caves with teeny keyholes looking out into the Net," Smarr said. "How can we go and get things when it's built on an unpredictable, shared Internet?" The academic community and private sector are solving the problem with private, high-speed fiber optic circuits (LambdaGrids) for about the price of a graduate student, instead of million dollars a month, according to Smarr. Compared to shared Internet connections, LambdaGrids are lower cost-per-unit bandwidth and don't have jitters, Smarr said.  The National Science Foundation OptIPuter uses LambdaGrids to visualize massive gigabyte data objects remotely, such as undersea hydrothermal vents. The National LambdaRail Partnership provides 40 10-gigabyte lanes for use in research applications.   

Most of the high performance grid deployments are in Europe and come from academia. "It seems like a silent revolution, but we've seen this at least two or three times before and it ends up with a major change," Smarr said.  He should know. Smarr was director of the National Center for Supercomputing Applications (NCSA) when in 1992 Marc Andreesen and colleagues developed the first Web browser--NCSA Mosaic.  

Topic: Outsourcing

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

1 comment
Log in or register to join the discussion
  • Grid your teeth

    Grid ain't happening! In order to roll out a Grid infrastructure, you need a homogenous computer namespace (i.e. all applications and data go into the same "buckets"). This does not exist at Company "F", so we are at a loss to implement Grid (with our 15,000 UNIX and 160,000 Windoze machines).

    New technology like virtual servers and the IBM P5 "throttling" chips help to drive utilization up - which is what Grid is supposed to do. This leaves little room for Grid. Combine that with thin clients, and Grid has NOTHING.
    Roger Ramjet