Say Goodbye, EDI

By David S. Linthicum



While we continue to struggle with the notion of business-to-business application integration, we are looking toward momentum technologies, such as message brokers, application servers, XML, and the Internet to provide us with the mechanisms for making B2B happen.

In fact, these days I rarely hear about more traditional B2B technologies, such as EDI (Electronic Data Interchange) which has been automating supply chains for years. Why not? It's because EDI is experiencing the death of a thousand cuts-and XML and the Internet are doing the cutting.

So, what happened to the interest in EDI considering the explosion in the B2B space? It's really a matter of the "three strikes, you're out" principle of the technology market.

Strike one, EDI is far too expensive. Ever try to enable your existing applications and databases to communicate with other systems using EDI? You're typically going to spend millions on technology and labor, not to mention the time-to-market cost. You'll need many costly items, such as data mappers to handle the differences in application semantics, a commerce server to control the information flow, and something called a value added network (VAN) to send and receive your EDI messages securely.

Strike two, EDI is not consistent. Let's face it: EDI is fundamentally flawed. Each EDI solution is unique for each trading partner, and they are unable to communicate without a lot of expensive and time-consuming coordination. If you think you can simply externalize application information as an EDI message and have that message processed by any other EDI-enabled system, you're dreaming. There are many versions or takes on the X.12 and EDIfact standard, and "proprietary" implementations abound.

Strike three, EDI is not positioned for the Internet. Simply put, the big EDI players including Sterling Commerce and Harbinger did not retool for the Internet in time, and as a result they're losing out to other XML players. While there are many efforts under way to marry EDI to the Internet, it's probably a case of too little, too late.

Even the EDI guys will tell you that EDI is inflexible, non-compressible, and costly, with its foundation more in the proprietary technology of the '80s than the open source and Internet culture of the '90s.

Underwhelming support
Of course, there is the issue of mind share. While many larger global 2000 companies have spent their way to EDI success, most midmarket companies find that EDI technology is out of their reach for the reasons I've already mentioned. In fact, only about 125,000 organizations worldwide provide EDI facilities to their trading partners, and that's according to the XML/EDI Group, a consortium of companies and individuals looking to merge XML with EDI. The United States tracks with that, with only 80,000 EDI-enabled businesses, which is a small percentage of those businesses that would benefit from B2B information interchange.

So, if EDI ain't making it, what's an organization looking to exchange information with other organizations to do? Clearly the Internet and XML provide the most flexible, uncomplicated, and open solution set.

Easy data movement
XML provides a simple mechanism to represent data from anywhere with a common point-of-integration for most business systems. The result is the easy movement of information between applications such as purchasing or shipping systems. Data moves with ease, since it flows within a format that's independent of the application and particular application semantics.

Also, XML's structure is easily adapted to support those found in EDI. Considering this, there is no need to redefine the meaning of particular commerce transactions (e.g., a purchase order), since XML is able support the same schemas supported by EDI.

What's more, XML is built from the ground up to support both the Internet and open software. XML parsers are free for the downloading, and B2B XML middleware, such as that offered by webMethods and NetFish, are much easier to use than traditional EDI processing environments. These new XML-enabled players provide EDI-to-XML translation services as well, giving you the best of both worlds and not forcing you to throw away your current EDI installations.

What's relevant about the rise of XML is the fact that the barriers to entry are slight. Where a traditional EDI-enablement project typically ran into millions, and thus was the exclusive domain of the larger players, XML levels the playing field providing B2B opportunities for companies on tight budgets. This makes XML, at least in my book, the great equalizer in the B2B space. But most larger companies including General Motors—whether or not they use EDI—have plotted a course to XML for future inter-enterprise application integration as well.

As a result of this XML mania, I'm seeing the EDI vendors who once laughed at the Internet as an unsure bastion of hacking hippies, now working hard to support it, and XML. For instance, the CommerceNet consortium is promoting the shift from traditional EDI, and the use of VANs, to Internet-delivered EDI, and they expect this EDI/Internet solution to run about one-tenth of the cost of a VAN. The CommerceNet standards rely on Pretty Good Privacy (PGP) and Secure/Multipurpose Internet Mail Extension (S/MIME) to transport EDI objects and keep the information private. Moreover, CommerceNet has announced the release of the eCo Interoperability Framework Specification, which describes how companies define their businesses via XML, as well as how EDI and XML should work together.

Linking XML and EDI
But the more aggressive XML/EDI efforts are coming from vendors such as NetFish and OnDisplay. These guys are both working on linking XML and EDI, as well as using the Internet for transport. The NetFish solution, called XDI, is able to create workflows along the supply chain as well as create XML-based forms that look like EDI. OnDisplay has released a similar product called CenterStage.

Yes, I know XML does not yet provide the standard security of EDI, but that's coming. I also understand that EDI has been in the B2B space for 20 years, and has many predefined information exchange entities (e.g., purchase order). But that doesn't change my deathwatch for EDI-at least from the point of mind-share, which is critical. EDI is joining Smalltalk, object databases, and the Apple Lisa on my list.