Commentary: The dawning of Sun's commercial Web software

Suppliers of Web-oriented system and tools software continue to play market leapfrog, with Sun Microsystems preparing to become the latest entrant early next month.

This is normal in any rapidly

	See news story: Sun sambas into Web-based software field with "Brazil"

maturing market. Existing products are renamed. Evolving technologies are bundled in new ways. Huge promises are made.

But inevitably, issues such as delivery, quality, functionality, compatibility, usability and profitability win out. The question is whether Sun, with this announcement, is willing to begin acting as a software vendor and not as a hardware vendor looking to drive hardware share with cool, but captive, software.

Sun has failed in most of its efforts to become a force in commercial enterprise software. With its announcement a month away and only fragmentary information available, it is too early to say what Sun's message will be. However, we can define a set of criteria that Sun must meet for its plan to be taken seriously.

First, its technology must encompass multiple applications--specifically, Windows NT and 2000 Server--and not be confined to Solaris.

To hold significant long-term interest to consumers, Sun must present a strong technology and strategy for becoming the first or second player on Windows servers after Microsoft. If it does not, it will lock both itself and independent software vendors (ISVs) out of the largest segment of the commercial marketplace. No one should view Sun as a software company so long as its software strategy remains devoted to driving its hardware market share.

Brazil's role
Second, this new application is built around Sun's Brazil, which has been available for experimentation as a free download from Sun's Web site since the third quarter of last year, Sun must present a strong plan for turning Brazil technology into product. This means the company must package Brazil with a full set of servers, including directory and application servers, and present a coherent plan for pushing the product through multiple channels. No organization is going to base its Web applications suite on an experimental product.

Third, Sun must specifically announce an inclusive list of protocols this application will support.

If Sun wants consumers to treat it as a software vendor, then it must offer the range of interoperability and compatibility that customers require. It must put leading protocols and methods on an equal footing and not confine itself to the electronic-business Extensible Markup Language (ebXML), its preferred Web protocol stack, which it is pushing to combat Microsoft. Even if Sun can't incorporate all the leading protocols immediately, it needs to show how it will be able to include them all within a year.

Fourth, Sun must explain how Brazil will work with Java technology, and it must commit to the ongoing investments necessary to keep the two coherently integrated as both are developed further.

The fundamental difference between the technology future espoused in the .Net philosophy and that touted by the Java world is a difference in the means by which services and interfaces are exposed to the outside world. Java continues with an application programming interface (API)-based approach, which is by definition tied to a complex implementation structure with various language bindings that need to be created. In contrast, .Net philosophically assumes all interfaces will be asynchronous and XML message-based.

This philosophy makes it much easier to gradually increase the number of services and to claim support across a broader set of applications and systems. In this way, a service may be consumed by software running on a given application, even though the instantiation of that service is not supported in that environment, because XML messaging (the interface model) is presumed to be universal. In the API model, an instance of that API (if not the service) must exist in the particular environment.

The suite hereafter
Fifth, Sun needs much more than Brazil alone to create a viable technology. It must show a suite of tools built around Brazil. Some of the additional technology it needs is in the hands of Netscape. However, because it is a separate organization in a separate location, Sun has experienced challenges in coordinating the technologies of the two companies.

Sun does not have many of the commercial components that it needs or has not established its versions of those components and services as serious contenders in the marketplace. Integration services is a good example of where Sun does have some technology inherited from the Forte acquisition, but this has not been broadly accepted as an integration solution in the way that competing offerings from, say, IBM have been.

Sixth, Sun also must have large ISVs on the speaker platform with it, pledging their support.

This must be a consumer-oriented presentation, not just a Microsoft-bashing session. If the only support for Sun comes from a couple of anti-Microsoft ISVs, consumers will have to question whether this is a serious response to real, and critically important, application user requirements.

With a crowd of major vendors--Microsoft, Oracle, IBM, Hewlett-Packard and now Sun--announcing plans for creating comprehensive Web applications, consumers may become confused. They must remember that first none of these applications are fully implemented and available yet. Consumers also need to avoid being swayed by competing vendor claims and concentrate instead on how the applications they run their business on can and will integrate with each other.

That is the test of whether any of these environments are viable and worth investing in.

Meta Group analysts Daniel Sholler, Peter Burris, William Zachmann, Dale Kutnick and Jack Gold contributed to this report.

Visit Metagroup.com for more analysis of key IT and e-business issues.

Entire contents, Copyright ? 2001 Meta Group, Inc. All rights reserved.