ASPs: The grass is always greener

By Neal Weinberg
Network World, 01/10/00

What do Amdahl, Dell and Deloitte Consulting have in common?

Don't know? Okay, what do Oracle, Qwest and Novell have in common?

Still stumped? How about IBM, Nortel and Cable & Wireless? How about Cisco, Citrix and Sprint?

Give up? The answer is that they all have an application service provider (ASP) strategy.

In fact, you'd be hard pressed to find a vendor without an ASP strategy these days. And we're not even talking about the whole new breed of "pure" ASPs like Corio, Applicast and U.S. Internetworking.

Which leads to my problem: If everyone is an ASP, how do the vendors differentiate themselves? More importantly, how are network executives expected to recognize the players without a scorecard?

Even if you have a scorecard, there are still some fundamental questions that need to be asked about this whole application rental craze.

First, let's assume that the ASP model makes sense for an Internet start-up with no IT staff, no legacy applications and no time to wait. And let's assume the ASP model also makes some sense for small to midsize businesses with limited IT staffs, limited technical know-how and limited interest in owning its own software.

How does that relate to a Fortune 500 company with a large, highly trained, experienced IT staff? A staff that plays an integral role not just in hosting applications, but in writing applications specifically for the unique requirements of that particular business, customizing applications to fit the specific needs of the business, and managing the performance of multiple, intertwined applications across LANs and WANs?

Another point: It's one thing if you're a start-up or a small company and you're buying your first software package in a particular area, such as customer resource management. Renting makes sense because you avoid the up-front investment. But if you're a Fortune 500 company that already has a full suite of applications - all of them paid for - what's your incentive to ditch an application that you already own? And if you can keep that application running for as long as you want, why would you exchange it for getting on the treadmill of monthly fees from now until the next millennium?

Here's another one. Let's say you're moving toward a directory-enabled, application aware, Quality-of-Service (QoS) capable, converged network. You want end users to be able to log on to the network from any location, enter a password and get access to their own specific set of applications. You want to set policies and have the network decide which applications get priority. You want to run voice, data and video across your all-IP network.

How does that effort square with offloading a particular application - say an enterprise resource planning application - to an ASP? Will the ASP take over your directories along with the application? Will the ASP make all the adds, moves and changes to your directories? Will the ASP handle all your QoS needs?

What if you have service-level agreements with various internal departments that require you to meet certain performance levels? Say you monitor those agreements with sophisticated network monitoring tools. How do you plug an ASP into that equation?

More questions: Will a single ASP take on all of the applications of a particular company? Or will an enterprise have relationships with multiple ASPs, with each ASP specializing in a particular application? In that case, what happens when an enterprise has relationships with 11 different ASPs?

Finally, and perhaps most importantly: The ASP model is built on a premise that applications will be delivered over the Internet. But even the head of the ASP Consortium concedes that companies won't rent applications over the Internet until reliability and security concerns are met. And when will that be?

Will the ASPs be able to answer these questions? I'm sure they're going to have to, if they plan to win business from the enterprise customer.