Build Systems, Not Companies, on Open Standards

Printer Friendly Version

Standards make possible the mass assembly of complex things, and many modern industries would not exist, at least not at their present scales, without binding agreements as to how parts should assemble into wholes. Imagine being a designer of electric shavers or refrigerators without the foreknowledge that 120-volt, 60/50 Hz electric power were available in target households. Although it might be possible to engineer custom products for customers with well-defined power supplies, the result would be more like an artisan's guild than today's mass market in consumer electronics. Unit costs would skyrocket, volumes would dwindle. The same would be true of an automobile market without commodity fuels available on every street corner, or of telephone service absent reliable wiring and signaling standards. Without standards, every job is a one-off, and vendor lock-in is assured after the first project.

This was more or less the state of computer networking before the Internet opened to business traffic in 1992. At that point, proprietary network protocols dominated, notably Novell NetWare between Intel-based PCs, Appleshare between Macs, DECNet between VAX minicomputers, IBM SNA between mainframes and several others. Into this oligarchy the Internet dropped a bombshell: TCP/IP, an open, royalty-free protocol stack that enabled any participating computer to talk to any other, not just on the company LAN but anywhere in the world. More, this capability brought with it a number of free client/server applications, including FTP (file transfer), SMTP (e-mail), NNTP (newsgroups) and HTTP (the Web). Because anyone could develop and embed these applications without paying license fees, they disseminated widely and, tempered in the fires of a friction-free market, rapidly approached a kind of bug-free feature-completeness rarely seen in commercial software.

When this happened in 1995-96 it was tremendously exciting and empowering. Today, it is easy to forget that open standards are still the beating heart of e-Business, without which concepts like intranets and business-to-business (B2B) commerce would not be practical on a mass scale. Open Internet standards-including the XML family of protocols and languages-continue to dramatically transform the economics of software production and consumption.

The rub

Yet it is equally easy to forget that few companies have figured out how to make lasting money on these standards or the products that implement them. Today the TCP/IP stack, a piece of open-source code, lives on every desktop as part of the operating system, affording no value-adding or OEM licensing opportunities. The same is true of applications like e-mail clients and Web browsers. Nor is the commoditization of these entities tied to the dominance of any one vendor; Microsoft Windows, commercial Unixes and Linux desktops all offer a high level of Internet functionality as a baseline, not a line item, as demanded by the mass market.

The benefit to users of having commodity networking in their home and business machines is clear. What is harder to see is how IT suppliers can thrive in this regime without redefining their value propositions. It is troubling that, having watched first-generation Web standards like TCP/IP and HTTP melt into the woodworks as differentiators, a majority of vendors are still bent on basing their business models on open standards.

The problem with basing business strategy on open standards is that, like a fire stoked with home furnishings, it eventually leaves everyone homeless. A standards-based technology market can burn brightly in the short-term. At the inception of a promising new standard, when industry consortia are first convened and vendors vie for dominance, the sheer utility of the innovation looks like a value proposition. It is-in the same sense that electric power and automotive fuel standards have value. Successful standards are those that gain universal acceptance and lower system design and assembly costs for everyone, broadening niches into mass markets.

But in the long run, when the dust settles and everyone is in agreement, open standards offer no differentiation and erode competitive advantage along the axis that has been standardized. Survivors are forced to rethink their strategies and offer value in wholly new ways, as electricity and telecommunications providers have begun to, with checkered success, under deregulation.

Think ASCII and TCP/IP. Think Web browsers and servers. Think SQL and ODBC. All revolutionary once upon a time, all beneath notice today. No sane business would position around any of them. Yet that is precisely what many of today's Java-centric and XML-centric vendors are busily doing. Smaller suppliers focused on these standards are doomed to consolidation, while their business unit counterparts in diversified firms are setting up to be loss leaders. In some cases this has the veneer of strategy; Sun and IBM dropping the prices (to zero) of their respective Forte and Eclipse development tools comes to mind. Elsewhere the vision is more obtuse, for instance in Hewlett-Packard's cannibalizing its Bluestone middleware brand by offering its J2EE application server free ($5k for the "resilient," fault-tolerant edition).

Apparently, HP, Sun, IBM and others are willing to sacrifice top-line revenue around products that implement Java and XML standards. It is less clear where the cross-subsidies that generate profits will come from. An interesting counterexample is Microsoft, which has taken an uncharacteristically proactive stance towards open standards in pioneering Web Services technology for .NET. For Microsoft, the commodity nature of .NET technology is explicitly a cost advantage and deliberately decoupled from the value proposition. While the openness of SOAP and WSDL drive down .NET production and deployment costs, revenue comes from the sale of fine-grained value-added services (Hailstorm), tied to a controversial but formidable customer relationship model (Passport). The approach may be crass by Internet mores, but it has the distinction of currently being the only open standards play that puts business first.

Gordon Benett is a technology strategist with over 16 years experience analyzing, architecting and developing information systems. He is currently with Aberdeen Group (Boston, MA), where as a Senior Research Analyst he follows the Enterprise Java and Middleware markets. Gordon founded Intranet Journal in 1996 and remains a reader and contributing author. He welcomes your comments at gbenett@mediaone.net.

Printer Friendly Version