Morphing is such an applicable term for the information technology industry. Technology often doesn't change; it morphs into something else. For example, the Internet's influence is causing client/server computing to morph. Business application vendors are scrambling to Web-enable applications.
The short history of business applications shows that businesses have deployed applications as products or services. In the beginning, the prohibitive cost of mainframes and software forced many businesses to rely on "timesharing bureaus" to manage applications and data. In today's jargon, companies outsourced applications to external "shared service" providers. Many companies still manage applications such as payroll and human resources this way. The big drawback of this old paradigm is restricted access to the data because the service provider essentially owns the data.
The pendulum swung the other way as the cost of hardware and software decreased and buying application packages off the shelf and implementing them in house became viable. Applications became shrink-wrapped products, and a market grew to supply installation, maintenance, and customization. In today's client/server age, the drawback of this product paradigm is the cost of acquiring, maintaining, and supporting the server and client software.
As the Internet begins to morph the client/server paradigm, the pendulum will probably swing back, and business applications will again become services rather than products. Some vendors are pushing the pendulum back: E3 Associates (www.com/e3 or 770-424-0100) has its ProfitTrack service bureau, a sophisticated inventory management package that very large retailers use to maximize their profit on inventory. The service bureau approach extends the benefits of the software to small businesses that could not otherwise afford it. Small retailers can improve their inventory management by linking their point-of-sale terminals with a service bureau that runs ProfitTrack software.
Business application vendors and sophisticated systems integrators need to think hard about the impact that, say, self-service information processing and electronic commerce will have on whether they sell their wares as products or services. Self-service processing and electronic commerce give internal corporate users and external corporate partners access to business information from anywhere, at anytime, via a desktop Web browser. As a provider or consumer of that information, the user or partner does not know or care whether the servers managing that application and its data are on a corporate LAN, WAN, or intranet. If the data is secure, functionality is not restricted, and performance is acceptable, the location of the server is largely irrelevant.
The stage is set for the introduction of the indenet, an independent service provider's intranet that hosts a business application that many external subscribers can share. The indenet's application servers run the business application server software. They serve up functional applets or HTML-format data to the subscribers' Web-based browser clients, with appropriate access security. Indenet application servers are linked to "customer" corporate database servers on a secured corporate intranet. The software vendor or provider manages the application server software, but the corporation using the indenet service maintains ownership and controls its data.
In theory, the indenet customer needs to own no application software, only an Internet browser. The indenet customer purchases indenet access licenses that let it access its intranet data servers through the service provider's indenet application servers. This arrangement provides the ultimate in value-based pricing because service providers can invoice indenet customers purely on a usage basis, and indenet customers don't need to purchase application software user licenses.
No one in the subscriber's business, except perhaps for a few system administrators, needs any traditional business client software: All primary and casual application users work with the data exclusively through the interaction of their desktop Web browsers with the indenet application servers. Similarly, the vendors, customers, and other extended-enterprise partners of the indenet subscriber can have usage licenses to the indenet servers. This capability lets business partners participate in electronic commerce in its many forms, using the indenet subscriber's business system. The business partners can access the business system in the same way that the data owner does--through the indenet service provider.
Indenets are a terrific opportunity for application vendors to change their business model to become service providers instead of product providers. Indenets also establish a more logical division of responsibility for managing business systems: The respective owners, the software vendor, and the data creators and consumers maintain their software and business data.
Positioning the application vendor as the service provider that maintains the indenet makes great sense. The vendor wrote the software and has the best staff to keep it running and delivering optimum performance. The vendor can manage software revisions efficiently because it has to upgrade only its inhouse indenet servers, not hundreds or even thousands of distributed customer servers.
Meanwhile, indenet subscribers retain complete control over their business data because the database server with the stored data is on the internal corporate intranet. The business IS department can provide any other kind of application integration or data access needed from corporate desktops because the data and the data source is within its control. The Internet provides the means to use platform-independent browser software to enter and access indenet data.
The indenet customer does not have to purchase or maintain the server or the client accounting software; the customer has to maintain only a database server and a proxy (security) server that it probably maintains anyway for other non-indenet applications. The indenet vendor can change its revenue model from a product-and-maintenance basis to a subscription basis and reduce support overhead (possibly, by an order of magnitude), which could lead to a much healthier bottom line.
NT Server for Indenets
So what's this idea got to do with Windows NT? NT Server is a good candidate for running the indenet application servers. On the indenet provider's intranet, each indenet subscriber can share or be allocated its own discrete application server for security and performance reasons. The indenet provider and indenet customer can use NT-based Internet connectivity and security servers. This idea naturally makes one think the indenet is an ideal business for Microsoft. Think of the revenue potential of a Microsoft-managed indenet running, say, SAP's accounting applications. It would be a multibillion-dollar service market.
IBM, of course, might argue that an indenet is an ideal paradigm for mainframes or AS/400s. This argument has many advantages, but an NT solution lets the indenet provider start small, using inexpensive commodity hardware and software, and build up. Consequently, business application vendors (such as accounting or sales automation vendors) could pilot indenets quickly and test the concept's business potential with relatively little expense.
NT Server is an interesting vehicle for an indenet provider's intranet, but this solution may discourage the indenet customer's use of NT Workstation: The indenet business model is better suited for the network computer (NC) as the indenet customer's primary desktop, because users need only a Web browser to interact with the indenet applications. The NC is a cheaper and easier way to deliver indenet connectivity. This approach would benefit the NC concept's non-Microsoft backers, such as Oracle and Sun Microsystems. But in the end, users will access the indenet with whatever is already on the corporate desktop--the well-entrenched PC.
Of course, indenet providers and customers would depend on a core infrastructure, the Internet, that is not wholly in their control. But as standards emerge for transaction security, data encryption and compression, and point-to-point tunneling techniques, this dependence may not be such a problem. The real problem may be persuading businesses to adopt the indenet model and go back to the future with a new type of outsourcing.
Still, I predict that the first vendors who package viable indenet services will lead the way toward morphing client/ server computing into a new way of managing corporate business applications. This new way combines the benefits of outsourcing application maintenance with complete internal control of the business data asset.
Mark Joseph Edwards,