Sun Microsystems was one of the key beneficiaries of the dot-com revolution. In fact, it was the self-proclaimed “dot" in dot com. But that notoriety—as well as the earnings and stock valuation that went with it—disappeared with the rest of the market.
Sun’s most compelling long-term technology has become its systems software—primarily Java—and not the hardware that has driven most of its revenues in the past. As Sun struggles to redefine itself as a systems software company, it’s between a rock (Microsoft) and a hard place (IBM and Linux). In this article, I’ll look at two of Sun’s three key post-Java software initiatives and how they are positioned in the enterprise space. In my next column, I’ll discuss the third key initiative and the partner Sun will need to make it work.
First priority: Universal authentication
Sun Microsystems’ chief executive officer, Scott McNealy, and several corporate partners announced an alliance for handling a computer user’s online digital identity. Using an online digital identity, a computer user can share stored personal information and use the identity as a key to control access to that personal information by Web sites or service providers. Dubbed the Liberty Alliance Project, this is a clear response to the single-vendor approach of Microsoft’s Passport service, part of the My Services (formerly HailStorm) initiative.
Microsoft’s Passport service stores commonly requested information such as names, addresses, and credit card numbers in a central database that is managed and controlled by Microsoft. In past public appearances, Sun’s chief technology officer, Greg Papadapoulos, made direct references to the need for "a universal identity system that doesn't play favorites.” Microsoft has responded by announcing that it will allow other companies to use Passport for authentication only and allow the companies to maintain the personal information of their members.
Microsoft isn’t the only competitor in the universal authentication battle. The Free Software Foundation announced an initiative called DotGNU, designed to create a universal authentication system that competes with Passport. And AOL has announced its Magic Carpet initiative to provide authentication services for its users.
Given McNealy’s ability to line up a star-studded cast of corporate sponsors for Liberty (including heavyweights such as American Airlines, GM, Bank of America, and RSA Security, among others), it’s likely that even though Liberty is barely even a specification today, Microsoft, AOL, and others who want to play in the Internet authentication space will have to find ways to interoperate.
So where is this all going? The most likely end result is an Internet-wide implementation of Kerberos with the ability for services to interoperate as independent federations of users.
Corporations wanting to grant users from different federations access to their sites will have to be intimately familiar with how to manage the interoperability themselves until this idea gets through some kind of standards process (and that’s three to five years away). The other option is to begin building around one standard now and hope that the standard upon which you build will be able to interoperate in the future. Many Microsoft-centric companies are already making those bets with Passport. Others will likely have to wait until Liberty or Magic Carpet becomes usable sometime in 2002.
What do you think of the universal authentication plan?
Will consumers be receptive to these plans to store their personal information in a central database on the Web? What kinds of business applications are there for this plan? Or is this a gamble that no one will win? Send us your opinion of universal authentication plans.
Second priority: Create a viable desktop competitor
Sun recently released the beta version of its office productivity suite, StarOffice 6.0. StarOffice 6.0 is a product based on technology acquired from a German software development shop in 1999. It replaces the prior version (5.2) but retains its claim as the only multiplatform productivity suite running on the Solaris operating system, Linux, or Microsoft Windows. Sun also made the bold move to release the source code to the suite to allow companies to customize or enhance the product. Most importantly, they also priced the product very competitively—as a free download for personal use and with low-cost licensing for corporate use.
Due for general release in the first half of 2002, the new version supports XML file formats, higher-quality Microsoft Office import and export filters (including support for Office XP), better performance, and new international versions (including Korean, Japanese, and Chinese). The international support is especially important for countries like China, where the government is mandating the use of Linux (I understand they prefer Red Hat) for most government projects. Although WordPerfect is still available for Linux, StarOffice is the only fully functional suite available from a major manufacturer. Given the retail price of Office XP ($479), many companies may see the Windows versions of StarOffice as a more palatable upgrade from their older versions of Office 95 or Office 97.
But what is Sun’s end game here? StarOffice doesn’t generate significant licensing revenue for Sun. It only has the potential (so far unfulfilled) to put a dent in Microsoft’s licensing revenue for Office. Prior versions of StarOffice have failed to even show up on Microsoft’s radar screen for three major reasons: performance, file compatibility, and programmability support. Version 6.0 makes great strides in the first two areas, but it doesn’t even attempt to fix the third. Unfortunately for Sun, many large corporations standardized on Microsoft Office because of their ability to use the macro language (VBA) to create templates and applications that could be locked down. As hard as it was for Microsoft Word to displace WordPerfect because of those users’ affinity for keyboard shortcuts and templates, Sun’s job to displace both Word and Excel will be even more formidable. The real opportunity is in the small to midsize enterprise.
But in the small to midsize enterprise, Sun has no presence on the desktop or the server. The only possible near-term outcome of StarOffice promotion in this market is the entrenchment of Linux servers and desktops. It’s unlikely that a small enterprise would adopt StarOffice on its desktops and then purchase a Sun server and Solaris server license for its office. When the ASP market was white-hot, many Sun partners thought the hosted StarOffice play would be their golden goose. But nobody, not even Microsoft, was able to create a compelling, hosted productivity offering.
StarOffice may yet be Sun’s best chance at capturing the desktop from Microsoft and continuing its conversion from a hardware company to a software company. But Sun won’t be able to pull it off by itself. Its likely partner also has a say in whether Sun's third key initiative will ever get off the ground. Next week, we’ll look at this initiative and that partner and how they are likely to work together.