Since electronic data interchange (EDI ) has earned a reputation as a complex, rigid, and expensive means of business document and data exchange among trading partners, one would expect it to be relegated as a relic of a bygone era. In the computer business, where new technologies can come and go almost overnight, EDI should have long become an artifact, let alone a technology with a mid-life crisis considering a number of younger technologies, such as DOS or FORTRAN, have already succumbed to extinction.
emerged in the 1960s when the railroad industry sought a way to speed up and
automate business communications between remote computer systems and to eliminate
the high cost of sending paper documents by snail mail. However, the concept
did not fully take hold industry wide until the 1980s when standards were introduced
to define data exchange and when it became apparent that business-to-business
(B2B) computer-mediated communications would require all parties to adopt a
common protocol for purchase orders, advance shipping notices (ASN), and other
pertinent documents. As a result, several technical committees have developed
protocols governing such exchanges, which channeled through value-added
networks (VANs) carrying these exchanges became collectively known as EDI.
finance, insurance, and other industries have heavily leveraged EDI and proprietary
communications lines to conduct business. Also, major manufacturers, such as
automotive original equipment manufacturers (OEMs) and consumer
packaged goods (CPG) companies, have embraced EDI and mandated that their
suppliers do the same. However, implementation of the technology has initially
caused many a headache particularly in smaller companies.
EDI vs XML
Indeed, EDI has a reputation for being expensive to set up and run, given the companies had to deploy numerous VANs, which drove their suppliers' network costs up and forced suppliers to be proficient in various communications protocols. Aside from the upfront cost of the EDI infrastructure software, companies choosing to go with a provider of VAN EDI networks, face additional costs in maintenance and transaction processing fees alone. For the above reasons, one would expect companies to begin gravitating in droves toward extensible markup language (XML) and related web services for transaction communications.
Indeed, in theory, XML shows many advantages, since unlike EDI, it was specifically designed to use the Internet as the data transfer mechanism. Also, while organizations in an EDI network have to set up direct point-to-point connections between each participating system, XML's extensibility means that any number of participating companies, which have agreed on a data format for transactions, can supposedly freely exchange data electronically. In short, XML has initially not only promised to ease the technical pain of integrating the flow of data between systems, applications, and people, but also to reduce the cost of this process through faster transactions throughput, improved trading partners' data quality, elimination of manual processes and, other potential cost saving benefits.
Nevertheless, while at the surface there would be few economic or strategic reasons for organizations to persist with EDI, many seem reluctant to adopt the alternative at this stage. In fact, there has been almost negligible growth in the number of organizations replacing their EDI-based systems with XML. Furthermore, it seems that sales of EDI-based products and services are even growing, given some estimates that EDI transaction volumes increased almost 20 percent in 2003. To be fair, though coming from a miniscule install base, XML transactions volume almost doubled during the same period and it still constitutes a single digit percentage of overall B2B transaction flows.
natural question is how come EDI has sustained its popularity and what factors
are inhibiting the take-up of XML in earnest. Well, it appears that there is
no immediate incentive for enterprises to move away from EDI. The key reason
for this is that the number of organizations using XML has not yet reached the
"critical mass" of at least double digits percentage of overall B2B data flow,
while the estimated number of large and mid-sized organizations using EDI is
estimated between 250,000 and 350,000 worldwide. For the time being, the wealth
of businesses that have invested significant resources in EDI still use the
technology for B2B communications, and many see EDI as the best choice for secure,
reliable transactions, given it is a mature, standardized, and trusted medium.
Namely, the leading EDI standards, such as the X12 and EDIFACT, continue to
meet the ever-evolving needs of more than a dozen industries, as opposed to
the ongoing evolution of a myriad of industry-specific extensions (dialects)
that are added to the base XML standard—many of which are non-interoperable
and still a work in progress. It is thus a small wonder that XML adoption and
perceived value is higher where there is a mature industry initiative and standard
like RosettaNet, CIDX, UCCNet,
AS2, and so on within respective hi-tech, chemical, and retail and consumer
traditional VAN companies have begun to do EDI business over the Internet and
the big buying giants that have always used VANs, such as Target
and Home Depot, are still very happy with the way that they
do business, because it works for them. And as long as these big hubs demand
their suppliers to do business with them via EDI, VAN providers are going to
stay in business.
when a large number of an enterprise's supply chain partners start implementing
XML will other trading partners start seriously considering XML as an alternative
to EDI. Particularly in the ongoing sluggish economic climate, companies are
typically looking to extract maximum value from their existing IT assets. As
such, C-level executives are highly loath to write off significant investments
in their EDI infrastructure, although some of these systems may date back as
far as the 1980s. This is particularly true when many vendors have so far failed
to make a compelling return on investment (ROI) case for a switch to
XML, because nowadays XML capital and labor expenses are still higher than those
of EDI since users need software (and accompanying skills) to do XML translation.
Conversely, despite high initial set-up and value added service costs, EDI implementations
can actually be cost-effective, while the advent of web-based EDI connectivity
standards, software, and services has lowered the entry barriers for many companies
that, initially, would not have considered it an option.
Namely, during the past few years, a number of EDI suppliers have breathed new life into this old workhorse technology by developing offerings that use the Internet as the communications medium, eliminating the need for multiple VANs and driving the per-transaction cost downward. In addition, vendors have developed hosting services that reduce or eliminate customers' need for in-house EDI resources, whereby, further, EDI complexities can be hidden behind a browser-based "thin" client interface, making EDI communications practical for many small companies that would not have considered its use before. As a result, although XML is more flexible and easier to use than EDI, EDI protocols will run for years among major manufacturers or retailers that have heavily invested in their use and have the clout to demand their trading partners use EDI as well. As a matter of fact, EDI, XML, and any other format are merely "semantics" and input streams for expressing data, and whether a transaction is transmitted in EDI format or XML is largely secondary to the fact that electronic document and data exchange is growing rapidly.
EDI, XML in Tandem
positive news for end users is that EDI and XML might not present an either-or
choice at all, but rather offers technologies in tandem, given that
significant interoperability between EDI-based and XML-based systems is already
taking place. The Internet Engineering Task Force (IETF), an industry
standards organization, has for some time been working on proposals for standardizing
a way to secure EDI transactions over the Internet. Despite this co-existence,
in the long term it might make economical sense to switch to XML, since the
plethora of market-driving vendors like Microsoft, Oracle
or IBM that have been incorporating support for XML in their
products as the enabler for web services (see Liberty Alliance
vs. WS-I; J2EE vs. .NET; Overwhelmed .YET? ) means B2B trading environments
will, ultimately, be dominated by XML. Furthermore, the price of many XML-based
software products will eventually come down, which is unlikely to occur with
EDI VANs, while the industry-specific standards will inevitably mature. Eventually,
the XML traffic will exceed EDI X12 protocol traffic, but it will not necessarily
be a replacement. X12 is still the dominant format for things like purchase
orders and invoices, whereas, for newer things like collaborative forecasting,
there are no widely used X12 documents so far and this might motivate users
to exchange data in XML.
is only logical that EDI VANs and other providers of integration as a service
would seek to reinvigorate their business value proposition by adding applications
to their "plumbing" portfolios to offer more of an application-like, vertical
solutions-oriented approach to trading community integration requirements. As
an example, vendors that have eyed delivering packaged data synchronization
software and transaction delivery services would be e-commerce network services
vendor Transora and product information management software
maker Trigo Technologies. Both these companies recently announced
plans to work together on a joint offering that combines their products.
contrary to the above competitors as well as to Inovis and
Sterling Commerce, which partnered to fill in key functional
gaps in product information management (PIM) and data synchronization
areas, the direct acquisition of HAHT by Global eXchange Services
(GXS) shows its commitment to the retail sector.
A company that is dependent on EDI for transaction routing should look hard at companies such as GXS as its pathway into the XML century. What is most important to them should be how smoothly the translation service can be implemented in terms of integration with their other systems. Therefore, size and technical strength are not as important as experience with the same systems users might have. But even if a company has the specific experience users need, it does not necessarily follow that they will benefit from it. Therefore, when negotiating terms potential users should make sure they have guaranteed access to the individuals who have worked on their kinds of systems. Given XML and EDI concurrent use for some time in the future, companies should think carefully about how to leverage the mix as to minimize the risk of duplicating efforts where both systems doing virtually the same work.