Many of my clients are getting serious about the Internet. Let me explain: No, they're not Rip Van Winkles who've slept through the technology innovations of the past 20 years. But they are thinking about how they can use Internet services to augment -- or even replace -- their WANs.

They have a lot of reasons. Some companies feel they're too small to deploy a full-fledged WAN. Others find the rates for Internet bandwidth are superior to those they can negotiate for traditional WAN services (such as frame relay and MPLS). A common problem is geographic footprint, particularly for access: Although a carrier serves a particular region, it doesn't provide local connectivity to the specific city or building needed.

In addition, there is a handful who simply wonder, "why not?" Internet reliability has increased steadily over the past several years. Consumer adoption of such applications as VoIP and video streaming would seem to indicate the Internet is fully capable of handling such applications. Moreover, application-layer security can do a fine job protecting sensitive data -- making a private network moot. So, many folks are beginning to wonder seriously why they should bother with a WAN at all.

Here's why eliminating the WAN entirely is a dangerous strategy. First, despite happy consumers touting VoIP services from such firms as Vonage and Skype, quality in enterprise deployments has varied wildly. Yes, I've spoken with folks who use, say, Skype to communicate with their outposts in Eastern Europe -- and who swear it works just fine. For every one of them, however, there are at least five others who say its quality is so iffy it's simply not enterprise ready.

Thus, although relying on the Internet as infrastructure for a certain set of applications (linking users to e-mail servers or Web 2.0 applications) can work just fine, relying on an Internet-based WAN can preclude some of the more innovative applications (such as VoIP and video over IP).

Second, the general perception that Internet services are cheaper than WAN services may be incorrect. Yes, the price of Internet access generally compares well to the price of services like frame relay and even MPLS. As carriers roll out low-cost Ethernet services, however, that's changing. A company that neglects carrier Ethernet in favor of public Internet services may end up paying more.

Finally, there's the long-term impact of a wholesale shift to the Internet. As I've noted previously, carriers don't make money from Internet services. So, in general they try hard to limit the use of their networks with such tactics as limiting bandwidth-intensive applications and handing off as much traffic as possible at the nearest peering point. The impact on users? Degraded service quality. So-called net-neutrality regulations won't make this better: In fact, they're very likely to make it worse, by ensuring that Internet traffic continues to be unprofitable.

What should companies do instead? By all means, augment the WAN with Internet and IP VPN services for hard-to-reach branch offices. To augment footprint, look to aggregators, such as Vanco Group (just acquired by Reliance Globalcom) and Virtela Communications. And consider carrier Ethernet to keep prices down. But don't ditch the WAN just yet.