Why the network will be the next computer

NetWorld+Interop 2002 clarified the industry's push toward a global, unified network infrastructure, based on Internet protocols. The writing is on the wall: Go IP or you'll wind up on a dead end

For the last few years the NetWorld+Interop confab fell out of favor. It lost the buzz native to events that help define the industry's new inflection points.

This year the atmosphere is different. The "cool" Internet of pervasive e-commerce and e-marketplaces eclipsing the brick-and-mortar world has passed into history for now. Instead, the Internet has returned to its roots (which were first exposed at Interop conferences in the 1990s) as core network infrastructure and applications platform, and grown way beyond its heritage in academic circles.

In fact, we are truly at an inflection point, bridging into the next phase of the Internet. We will be able to look back at this year's N+I and say we saw not just a few indications of an economic recovery, but signs of a future in which the network truly is the computer, to borrow a phrase from Sun Microsystems.

The inflection point in evidence at N+I is the push toward a global, unified network infrastructure, based on Internet protocols. The benefits are well articulated at this juncture in terms of cost savings and flexibility, as well as industry standards and support.

N+I keynote speakers Serge Tchuruk, CEO of Alcatel, and Cisco CEO John Chambers both identified interconnected IP-based LANs and WANs that move voice, data, and video as a key enabler for more cost effective and useful Web-based applications. Tchuruk termed this ultimate evolution of IP networking protocols and open standards as the "borderless enterprise. Chambers called it the "network virtual organisation."

Whatever you call this movement, it's more a question of when and how rather than if IP-networks will become the network of networks. "Almost no CIO I talk to today disagrees that within five years we will have a single infrastructure for data, voice, and video," Chambers said. The when and how is tied to providing migration paths that allow for more gradual replacement or upgrading of existing equipment within businesses.

Tchuruk said that enterprises don't need to take a "forklift" approach and replace legacy systems, but should be able to migrate to IP-based network services at their own pace. For example, deploying voice over IP (VOIP) can be done in combination with traditional phone services.

Vendors hope that this migration to a more IP-based solutions will catalyse spending and a return to profitability for their customers and themselves. In reality, unifying network architectures with Internet protocols is just a first step. Both Chambers and Tchuruk stressed that these networks must have carrier-class reliability, quality of service, and bulletproof security to succeed with enterprise customers and consumers.

Broadband and Web Services
Getting to the promised land will also require parallel developments in broadband and Web services. Let's take broadband, or bandwidth, first. For various reasons, such as cost and last-mile connections, bandwidth still has a way to go before it's more like water flowing from a faucet. Only about 20 percent of U.S. Internet users have broadband connections at home. The tough economy has put a damper on broadband. Alcatel accounts for a majority share of installed DSL lines, but saw the growth rate drop significantly last year. Tchuruk believes that the promise of real-time, multimedia applications will push the carriers and service providers to more aggressively build out the broadband infrastructure on top of IP-networks. "In a phase where carriers use every trick to delay investments, it is consumers who will drive demand," Tchuruk said. We have heard this story before. Consumers will drive more investment in broadband because they want video on demand on their handheld devices? So far, that's not the case. Clearly, entertainment will play a big role in catalysing a broadband buildout, but only if the entertainment industry stops trying to impose its last-century business models onto the Internet. Businesses are increasingly demanding anytime, anywhere access to rich data, but also view it through the ROI lens. A unified IP network is one step to enabling a mobile workforce, but fixed broadband wire and wireless buildouts are crucial. More intelligent traffic management and the move to IPv6 can also deliver more efficient use of network bandwidth. (See "The incredible shrinking Internet.") Web services is one of those phenomena, like the Internet itself, that has toiled in the shadows for years and suddenly jumped into the limelight. Now practically every product claims some Web service association (even if it doesn't fit the basic definition of Web services). While much progress has been made with standards like SOAP and WSDL, the hype has far surpassed reality. The natural inclination of companies is to create proprietary intellectual property that can be monetized at the expense of competitors. The path to consensus agreement on standards that benefit the many is a long and bumpy road. It will be a few years before the true shape of Web services will be established, especially those that take friction out of vertical industry supply chains, for example. Nonetheless, Web services under the .Net or Java platforms have serious momentum. The combination of Web services and IP networks is potentially the fabric from which applications will be woven. A unified IP network also paves the way for grid computing. I talked at N+I with IBM vice president Irving Wladwasky-Berger, who described grid computing as making the Internet a computing, not just networking, platform. "The Internet is a great network, communications mechanism and repository of content, Wladwasky-Berger said. "The next level is applications that can be distributed all over the Internet and accessing all the resources they need over the Internet." The scientific research community is already using grids to link supercomputers into a single virtual machine, using workload managers to distribute the jobs across the machines. Essential grid protocols are also in the works to create a standard framework of services, and eventually an open source, distributed operating system to work over the Internet. Many pieces have to fall into place almost in parallel for the next phase of the Internet to succeed, but the writing is on the wall. Go IP or you will end up on a dead end path.
Have your say instantly in the Tech Update forum. Find out what's where in the new Tech Update with our Guided Tour. Let the editors know what you think in the Mailroom.