BOSTON -- Are intelligent networks the wave of the future, or will the "dumb" network model that lacks embedded...
intelligence continue to serve as the foundation upon which networks are built?
That was the question put before two industry experts during a lively debate last week at the 2004 Next Generation Networks conference.
David Isenberg, a consultant with Cos Cob, Conn.-based Isen.com, argued that dumb networks, which simply transport packets of data, provide end-users with the most control over network applications. By leaving intelligence at the network edge, Isenberg said dumb networks also encourage greater innovation because the dumb network is always open for new applications.
"End-users want physical redundancy, real competition and control at the edge," Isenberg said. "Stupid networks allow for smart applications at the end-point." As an example, Isenberg cited the recent success of Skype, an IP telephony service that offers high quality voice calling directly over the Internet, the ultimate dumb network.
So even though dumb networks may not be as reliable at the physical layer as smart networks, they are "reliable where it matters…at the application level," Isenberg said.
Taking the side of the smart network was Tom Nolle, president of consultancy CIMI Corp., headquartered in Voorhees, N.J. He argued that the smart network model offers the greater potential for technical innovation.
Because an intelligent or service-specific network is inherently more valuable than a dumb network, providers of smart networks are able to charge higher prices for their services, increasing profit. It is this profit, reasoned Nolle, that will be pumped back into the development pipeline, thus ultimately spurring more and greater innovation.
Nolle contrasted this argument by explaining the basic business model associated with dumb networks, a business model that he believes is doomed to failure. The reason for this, said Nolle, is that there simply isn't enough money to be made in the business of providing dumb networks.
"You cannot have a market of only consumers," said Nolle. "Providers are a necessary component."
Isenberg countered by pointing out the old business adage that "the customer is king." Dumb networks, he argued, can and will be profitable because if there is customer demand, providers will step up to fill the void.
Still furthering his case, Nolle contended that smart networks open up more applications to more people than dumb networks. Because smart networks allow applications to be simplified for the end-user, they enable more user-friendly applications. In today's business environment, this could prove to be more important than many realize.
"Ninety percent of businesses have no technical support at all," stated Nolle, meaning network management and other IT functions are being conducted by non-IT staff.
"If you have a really smart service," said Nolle, "it can be run by really dumb people."
Again, Isenberg disagreed, arguing that smart networks are simply not agile enough for the real world. "They're rigid, expensive and complicated," he said.
In the end, the only point that Isenberg and Nolle managed to agree on was that eventually there will be a clear winner. Both believe that smart networks and dumb networks won't be able to coexist for long.
As for the audience, they appeared to be just about as divided as the two speakers themselves. A show of hands evidenced a near 50-50 split among the attendees, with just a little more than half supporting the smart network position.
One such person was Donald Root, a product manager for Ultra Electronics - DNE Technologies. Prior to the debate, Root said he was solidly behind the dumb network model. After listening to the arguments, however, he had switched camps.
"Tom (Nolle) really changed my mind," he said. "His whole premise about the economics and ROI model (of the smart network) really made sense."