Standards for e-business take the spotlight in the battle for XML dominance.
For several years, the World Wide Web Consortium’s (WC3) Extensible Markup Language (XML) has been hyped as potent medicine for all the Web’s woes. Solutions to issues like interoperability, indexing, and searching for relevant information are within sight thanks to XML, but the success of XML will depend on various standards initiatives. How XML actually gets implemented and who will control those standards is at the heart of a fierce conflict.
In theory, XML is supposed to act like a universal business-language adapter, allowing companies with disparate ordering, invoicing, fulfillment, and other business processes to communicate without human intervention. With XML, business processes can be automated: For example, you can create purchase orders that are machine-readable. These documents can then be automatically routed to various partners with data fields intact.
In practice, XML has actually somewhat muddied business communications, as businesses adopt different XML standards that treat mission-critical business processes differently. Two partners that use different XML frameworks need yet another XML adapter to talk to each other.
Until a clear standard-bearer comes forward, businesses need to carefully examine the state of XML standards and choose wisely, keeping in mind what present and future partners will adopt. The two main competitors for de facto XML standards–Microsoft’s BizTalk Framework and e-business XML (ebXML), from an international consortium–both have their evangelists and detractors. BizTalk has the early lead because of Microsoft’s open approach. But ebXML is overseen by an independent, vendor-neutral standards body, which has advantages of its own. And there are several smaller players as well.
“XML adoption is spreading like wildfire,” says Brian Loesgen, principal software engineer for Stellcom, a San Diego-based wireless-integration engineering firm. Loesgen sees that businesses are hungry for XML and are adopting it right now. But for the time being, there is no clear e-business standard of choice.
Now’s the time small-to-midsized businesses are starting to plan XML integration projects with partners, clients, and suppliers. Businesses want to avoid creating additional layers of XML to interoperate with different XML frameworks adopted by their partners.
Promise and problems
The goal of XML is laudable: to create a system where all that data zipping across the Web becomes meaningful. You might recognize Albert Einstein as a name, but a Web browser just sees it as a string of text. But XML lets a company assign markup tags that give those letters a semantic meaning, like name or physicist. Consider your average business form–with data such as part numbers, dates, amounts, purchase order numbers, and the like–and it’s easy to see why XML is of significant interest to many companies. The Web, like a telephone, may allow different businesses to communicate, but they aren’t necessarily speaking the same language. XML allows the creation of a meaningful Web.
While XML provides a basis for exchanging documents and data over the Web, it’s not a complete solution on its own. “XML isn’t a panacea,” says Mort Rothstein, vice president of engineering for OrderFusion, a San Diego-based e-commerce platform developer. “It doesn’t do everything.”
XML isn’t so much a standard by itself as a system of rules for building other standards. Right now, there’s a flowering of XML standards across a variety of businesses, and the challenge for any company is implementing a system that will be around in the long run. “XML lends itself well to standards–that’s why there are so many of them,” Loesgen says.
XML essentially allows anyone to create a DTD (document type definition) or schema, so it’s relatively easy to develop a specification–a double-edged sword for companies looking for actual solutions. The threat to XML is that there will be too many competing standards, making it impossible for a business to choose one that will be compatible with its partners’ implementations. The success of XML in general, and any specific standard, depends on the development of particular implementations which are then broadly adopted.
The question of who will develop those standards cuts to the heart of XML. Standards tend to exist in a state of tension, pulled between twin forces. To be successful, any given standard needs to build in enough functionality to satisfy users, while not being bogged down by the process of hashing out that specification. Release a standard too quickly, and it likely will not cover enough ground to be useful. Take too long to develop a standard, and it’s obsolete before it’s released.
The tension that defines standards in general applies to XML in a special way. Because XML is more of a metalanguage than a language itself, it is easy to create standards that seem complete enough for B2B communication. Eager to get their standards out there, lots of vendors developed so-called standards independent of one another. In this case, most vendors released their standards too soon, before they would be truly useful for businesses. As the more useful standards–principally BizTalk and ebXML–grab market share away from lesser players, the situation will eventually work itself out. But it will take years.
Horizontal or vertical
Broadly speaking, there are two approaches to XML: horizontal standards (aimed across multiple industries) and vertical (specific to the business logic of a particular industry, such as manufacturing or retail). Vendors and industry groups are all developing and implementing their own e-business transaction standards, based on generic XML. But each of these standards is slightly different.
Different industries have different needs for transactions, hence different kinds of XML standards. Loesgen believes that industry participants will come together to develop a standard. “Groups are forming their own vocabulary,” Loesgen says. But the difficulty is hammering out a one-size-fits-all standard. “You never get something that accommodates everyone’s needs,” he adds.
With more participants in standards development, the need for diversity must be balanced against unwieldiness. Essentially, a compromise must be reached between a standard’s complexity and the speed and ease with which it can be developed and implemented. The risk is that standards development drags on slowly and eventually gives birth to a monstrously complex standard. “This is sort of what happened with EDI,” Loesgen says.
While there’s appeal to vertical standards–after all, who knows the business better?–it’s unlikely that there will be complete XML solutions developed by industry consortiums. “I doubt if you’ll see any vertical solutions from top to bottom,” says Will Zachmann, vice president of Stamford, Conn.-based Meta Group’s Server and Infrastructure group.
A strength of XML is the relative ease with which information can be parsed and transformed, so it’s likely that most companies will adopt hybrid solutions. But the foundation of e-business systems for a company will likely be based on broad standards, such as BizTalk or ebXML. “The overall direction is towards broadly shared horizontal standards,” Zachmann says.
According to OrderFusion’s Rothstein, specific industries will likely adapt their basic horizontal XML infrastructures to handle their specific needs. Most of the broad XML standards are designed to be flexible and modular. Says Rothstein: “ebXML has extrensics–a place to store things you didn’t think about.”
If industries are likely to develop specific solutions on top of horizontal technology bases, the question then becomes: Which basic infrastructure will survive? There are many XML e-business standards, including RosettaNet, developed by a global nonprofit e-business standards body, and Commerce XML (cXML) from Ariba, based in Mountain View, Calif. But the front-runners are BizTalk–developed by Microsoft, IBM, and a host of partners–and Sun Microsystems-backed ebXML. The latter standard is being developed by the Organization for the Advancement of Structured Information Standards (OASIS) and the United Nations Center for Trade Facilitation and Electronic Commerce, the body that developed EDI.
For the moment, Microsoft’s BizTalk seems to have the upper hand. The key to Microsoft’s perceived edge is Simple Object Access Protocol (SOAP), a network protocol that lets software objects from different languages communicate with each other. SOAP is key because it enables companies to use existing objects, beans, and other software components, and tie them all up in an XML framework. Even Sun has endorsed SOAP as a way to enhance the ability of Java to plug into existing enterprises.
“My sense is that the faction with the overall traction is Microsoft’s and IBM’s approach, built over SOAP,” Zachmann says. “The big guns have publicly aligned with SOAP.”
BizTalk also has an advantage because it’s closer to being reality than ebXML. Having the creators of EDI developing ebXML doesn’t stir confidence in Zachmann: “ebXML is the old EDI crowd, a bureaucratic crowd.”
Stellcom’s Loesgen agrees. “BizTalk will facilitate the adoption of XML. It’s well positioned to facilitate data interchange.” He sees SOAP becoming a de facto standard, thus greatly helping the case for BizTalk.
Usually accused of being tight-lipped on issues of proprietary standards, Microsoft is surprisingly open with BizTalk; the specification is well documented and easily available. “It’s a much more wide open process,” Zachmann says. “The specs are hanging out there for everyone to see.” In the game of vendors pushing for standard dominance, Sun is seen as being the one playing catch-up. “Sun is looking to be something of a Microsoft spoiler,” he adds. But Zachmann is quick to point out that momentum for BizTalk hardly means a death knell for ebXML. “It isn’t really an either-or choice.” he says. “I doubt that there will be two totally disconnected, incompatible standards.”
Zachmann believes that there will be levels of compatibility and overlap between the two standards, with individual businesses choosing a standard that makes the most sense for their business needs. “ebXML is promising better security,” he says. Additionally, ebXML extensively uses EDI specifications, making it attractive to companies already invested in EDI. “You could end up with specific pockets [of ebXML adoption], like the financial community, which is heavily into EDI,” he adds. “ebXML will gain some adherents, but they’re fighting an uphill battle.”
Rothstein says OASIS is trying to learn from the mistakes of EDI with ebXML. “ebXML pulled together the lessons learned and functionality from other standards, such as EDI,” he says. The question remains, though, whether ebXML can deliver on its potential in time.
According to Zachmann, “there’s huge momentum for the SOAP approach.”
While the battle for e-business infrastructure dominance rages, no one is questioning that XML is a vital technology. While EDI, the conceptual precursor to XML e-business implementations, enjoyed only limited use in specific industries, XML is seen as having a greater chance for long-term success, thanks to a number of factors.
“XML is succeeding where EDI failed because the barriers to entry are a lot lower,” Loesgen says.
If a company uses XML, it has already made certain decisions about data types and format. This greatly simplifies interoperability questions down the road. The competing XML standards and their shepherds are striving to avoid the pitfalls that plagued EDI–infighting and competing interests. Software vendors know that if they start going down EDI’s path, their efforts may be wasted. “Vendors need to make customers happy,” Rothstein says. “The standards world is getting smaller.”
This will be a big year for XML, as more businesses–especially those of the small and midsized variety–adopt XML-based e-business systems and figure out XML’s place in their strategies. XML and its underlying standards are still a work in progress, and the subject of which standard to implement is hotly debated. “There are a variety of firefights still to be held,” says Zachmann. But despite the debate, one thing seems clear. XML, in one form or another, is a business technology that’s here to stay. The imperative from businesses looking for e-business solutions has generated a groundswell of support for XML that seems likely to make it as ubiquitous as HTML.
As Loesgen says: “XML is EDI for the masses.”
Sean M. Dugan is a contributing editor for ComputerUser and InfoWorld magazines.