A battle is brewing for dominance of the standard middleware infrastructure for client/server computing. Ever since it became clear that the mainframe environment would not continue to be dominant, there has been confusion about what would replace it. During the heyday of the "open system" movement, it appeared that a consensus might form around products such as Open Software Foundation's (OSF's) Distributed Computing Environment (DCE). It is increasingly clear to me, however, that the de facto infrastructure will emerge out of a fierce battle for control. Who will win this battle? There is no simple answer.
In this two-part series, I suggest four possible scenarios for the future of the client/server infrastructure, and offer my assessment of each scenarios' probability of success. I discuss the first two this month and the final two next month. Each of these four scenarios will likely emerge over the next five years, and each has the potential to succeed. I'd also like to hear your opinions. Feel free to email me via the Internet at Jhurwitz@hurwitz.com.
The Microsoft Scenario. Microsoft (Redmond, Wash.) has amassed enough clout and mindshare to set the agenda for a future client/server infrastructure. The company successfully stalled the evolution of a standard middleware infrastructure for several years by promising a client/server version of Object Linking and Embedding (OLE). This gave Microsoft the time to rewrite OLE so that it would be appropriate as a large-scale client/server infrastructure. Microsoft uses the resources of the most successful, large software and integration companies, such as Digital Equipment Corp. and Software AG, to help it add the additional pieces it needs to complete an enterprise infrastructure. Microsoft is now poised to be the IBM of the early 21st century.
The United Front Scenario. Several large system vendors and partners, including systems integrators and large, independent software companies, come together and agree on certain infrastructure components that they all promise to use. These vendors are motivated by fear of domination by either the packaged software players or Microsoft. ISVs flock to this solution because it protects them from total dependence on either of these powerful forces.
The Internet-as-Infrastructure Scenario. The Internet, with its flat hierarchy and directory structure, forms the foundation for a new infrastructure. ISVs, sensing opportunity, add a variety of low-level services, such as tight security and development tools, to turn the Internet into a true client/server infrastructure. This is very appealing to customers because no one company prevails.
In addition, I am assuming that the company would use de facto and de jure standards to create both a remote procedure call (RPC)-based and messaging-based infrastructure, because support for both messaging and RPC would be necessary to meet a broad range of development requirements. This company would also have to create a comprehensive set of systems and applications management tools designed to help organizations master the deployment side of client/server computing.
Another assumption I would make for this scenario is important: The company must understand how to create a toolset that enables other software companies to build new applications that leverage the environment's infrastructure. If this packaged software vendor is good at partnering with other software companies, it will publish its APIs and make it extremely easy for others to add functionality. If enough influential corporations implement this software environment, they will also demand that their key vendors and partners use this infrastructure. Systems integrators will then sign on to provide service and support.
What are the ramifications if this scenario is successful? The most significant ramification is that there would be a profound impact on operating system vendors. If you assume that the packaged software company will provide a cross-platform environment, then operating systems, databases, and even systems platforms will take a back seat. This company would therefore be in a position to dictate standards in several different areas. SAP America Inc. (Philadelphia), with its Basis middleware layer, is one company that could conceivably pull this off.
Organizations that spent millions piloting early implementations of CORBA and DCE are now very familiar with their complexity and would love to turn the problem over to someone else. Microsoft has now bought itself several years to begin putting the pieces in place to take over the client/server infrastructure. Microsoft's first step is to rewrite OLE on the server into a comprehensive set of services that includes pieces of DCE (Microsoft has already copied its RPC) and components of CORBA to create what Microsoft refers to as Network OLE. The idea behind Network OLE is to create seamless links to desktop-based OLE. The company brings in partners eager to ride on its coattails to design the complex tools needed by most enterprises, such as security, performance monitoring, and so on.
In theory (and given enough time and money) this plan could work. Microsoft will probably purchase some technology companies that have expertise at the enterprise level. So, why do I give this scenario only a 40 percent chance of success? Essentially, I think that Microsoft will seriously underestimate the complexity and demands of getting the job done. It will come out with interim deliverables that will fall short of its promises. The complexity of distributed computing will most likely make corporations realize that a large-scale, general-purpose infrastructure is far too complex for any one company to implement and control.
What are the ramifications for the industry? If this scenario fails to materialize, the computer industry will relegate Microsoft to a lesser role than the software giant would like. However, Microsoft will clearly continue to dominate the desktop platform. I expect that the company will do exceedingly well in the LAN-based database market. In addition, Microsoft will continue to sell more and more component products that leverage its desktop franchise. This will open the door for more healthy competition and innovation. Next month, I'll discuss the final two scenarios: The United Front and the Internet-as-Infrastructure.