NW200: Nolle, 4/26/99

Sponsored by:

This story appeared on Network World Fusion at www.nwfusion.com/nw200/200end.html

Industry watcher Thomas Nolle shares his view on what would most revolutionize the network business and have the biggest impact on the future of our Network World 200 companies.

Free bandwidth -- what a concept! Nothing would revolutionize communications as much as it would.

Unfortunately, free bandwidth is impossible to reconcile as a money-making model for carriers. Fortunately, developments in networking are going to offer limitless bandwidth under terms users and carriers can live with.

These developments will change the world, making many of the things we now do without network support -- from meeting customers to buying custom music CDs -- possible on the network.

Recent advances in optical networking -- particularly dense wave division multiplexing (DWDM) -- make it clear that the cost of optical information transport is dropping to the "cents per megabit-per-second" level. But service providers have had no incentive to invest in this technology given the fact that it would lower service prices.

If bandwidth is free, nobody will want to sell it and so no one will invest in creating that free-bandwidth state, so the theory goes. The truth is more complicated. If bandwidth is free, nobody will want to sell bandwidth, but selling services based on bandwidth would be fine. We can make bandwidth free if free bandwidth is somehow walled off behind a barrier we call "services."

Service by proxy

The thing that makes a limitless-bandwidth network possible is a service proxy -- hardware or software that sits between a user and the Internet or other public data network and communicates with the user to secure network resources such as bandwidth.

A voice service proxy would draw bandwidth from the network's pool to support voice calls; it would draw on additional resources for calling features such as 800-number dialing. A CD-by-Internet proxy would draw on bandwidth to deliver content to a shopping mall kiosk at which a CD read/write drive would burn custom tracks for a buyer.

A service proxy can validate the user of the network service and limit the user's interactions with net resources. The service proxy makes it impossible to use CD-by-Internet bandwidth to support voice calling, so the cost of the new CD-by-Internet service can't drive down expenses and profits on voice calls. This protects the service provider's current revenue while creating new revenue opportunities. With the promise of new revenue, service providers can justify investment in DWDM technology, and the limitless-bandwidth concept makes economic sense.

Service proxies actually aren't a new idea. In fact, all the products aimed at providing voice services over the 'Net are service proxies. The products interact with a voice user using voice-calling procedures, then negotiate network connections and transport voice using Internet rules. A service proxy speaks the language of the network and the language of the user/service, acting as an interpreter between the two.

If service proxies aren't new, and if this idea is really workable, why don't we have CD-by-Internet today? There are two reasons.

The first is a service problem. Carriers already sell bandwidth in the form of T-1, T-3, OC-3 and even higher leased lines. And they've been steadily dropping prices to encourage users to buy more bandwidth.

But now, facility-based carriers such as AT&T, MCI WorldCom and Sprint have found themselves selling leased-line bandwidth more often to competitive carriers than to end users. If they maintained this "bandwidth to all" selling practice, they'd be building CD-by-Internet service opportunities for other carriers.

The second is a regulatory problem. The regional Bell operating companies are required to wholesale the elements of retail network services to competitive local exchange carriers (CLEC). This might make it hard to protect the limitless-bandwidth network's service prices; CLECs could simply acquire wholesale bandwidth and undercut RBOC prices.

Fortunately, these problems will be resolved within a year. Facility-based carriers have already stopped reducing the cost per bit on higher-bandwidth leased lines. Instead, carriers are moving users to frame relay, ATM or IP virtual private networks (VPN), offering service-level agreements rather than bandwidth. In short, these networks are based on the service proxy concept. And RBOCs are about to break themselves up into wholesale-only regulated divisions and competitive retail data and CLEC divisions. The latter will develop limitless-bandwidth applications, such as CD-by-Internet.

What's in store

First, service economies will increasingly favor value-added services rather than leased-line, bandwidth-only services. Users should expect to migrate to frame/cell or IP VPNs because the terms for those networks can be made favorable.

Second, users should expect service "filters" that will discourage or prevent the use of one service to support traffic of another. IP-based applications, for example, can be made less expensive only if the services aren't usable for voice transport. Other-wise, the service provider cuts its profits.

That's the real impact of the service proxy. A service proxy will be used to create specialized user-to-network relationships. The proxy will apply technical features to limit its use to the specific application it's designed to support. This will open the opportunity to deploy thousands of new network-based applications generating hundreds of billions of dollars per year in new revenue.

High-quality virtual conferences will be an alternative to air travel. And push technology could be used to distribute everything from breaking news to software updates. These have been suggested, but the current network business model has prevented their exploitation. The service proxy model will encourage these applications because carriers will make money, and that will spur massive investment in optical technology.

This shift will occur as soon as 2001 or as late as 2010. Understanding the truth will help make the earlier date a reality: A business never invests to lose money, so the only path to creating free bandwidth is the path that leads to carriers selling not bandwidth, but service.

Nolle is president of CIMI, a technology assessment firm in Voorhees, N.J. He can be reached at (609) 753-0004 or tnolle@cimicorp.com.