Can the problem of interoperability be addressed to make Web
services work for business? Users and suppliers have united to form
a body charged to make it possible
Last month saw an impressive gathering of the great and good in the
world of Web services - the technology that will allow applications
to interact over the Internet. Key suppliers such as IBM, Intel,
Microsoft and Oracle, joined forces with user giants Daimler
Chrysler, Ford Motor Company, Reed Elsevier, Reuters and United
Airlines to create the Web Services Interoperability Organisation
(WS-I).
Interoperability, or lack of it, is an issue that dogs many
promising technology developments. Since the IT industry seems to
have only recently addressed the problem of sharing word-processing
files across different systems, interoperability is vital if the
Web services concept is ever to produce practical benefits for
businesses.
IBM and Microsoft have worked for some time on a common model for
Web services interoperability (called the "Web Services
Framework"); the WS-I is their attempt to broaden support for their
effort.
Although Sun, which invented Web development language Java and
manufactures popular Web server technology, has been hesitant to
join the WS-I, the signs are that it will join in time - and this
can only be a good thing.
So what exactly does WS-I plan to do? The good news for users is
that WS-I is not aiming to define standards. The organisation's
members are clear that defining standards is the role of the
Internet Engineering Task Force, Organisation for the Advancement
of Structured Information, the World-Wide Web Consortium and
others.
WS-I will focus on ensuring the interoperability of suppliers'
implementations. This is fundamentally different in practice. WS-I
intends to do this by creating "profiles", which group together
sets of complementary standards - and using these profiles as the
benchmarks for testing.
The concept of working with profiles, rather than with individual
standards, is a sensible one, because they work at a higher level
that corresponds much more closely to the way that standards such
as Soap, WSDL (Web Services Description Language) and UDDI will be
used in real-world projects. The WS-I members are splitting into a
number of working groups, each of which will focus on building
sample Web services implementations which demonstrate the
technology at work in a truly interoperable, cross-supplier
environment.
This is where the value of having IT users in the organisation
comes to the fore. Without the input of leading-edge users such as
Ford, Reuters and United Airlines, the implementation scenarios
would probably be neither representative nor real-world - and the
WS-I would be just another talking shop. With their presence and
input, the sample implementations stand a good chance of helping to
educate other user companies, and software suppliers.
This is a great start - but we are looking to WS-I to do more.
Successful Web services will depend on the establishment of
vertical trading agreements within industries. They will also
require interoperable technology implementations within companies.
The profile mechanisms used by WS-I will provide an ideal vehicle -
but trading agreements will not be forged without the active
participation of representative user organisations to drive it. The
involvement of user organisations in WS-I is encouraging, as is the
membership of the standards consortium POSC from the energy
industry.
We would like to see much more widespread vertical market support
from enterprise users and other industry bodies, such as
RosettaNet, which was set up to standardise interfaces for
electronic commerce between supply chain partners. Given global
e-government initiatives and the role of regulation, it will be
interesting to see whether they become members. These players
should be at least closely monitoring WS-I's activities.
WS-I is a great addition to the fledgling Web services technology
industry. It is not just a noise machine, although we are a little
surprised and concerned that authentication and authorisation are
not referenced, despite the existence of Passport, the Liberty
Alliance and emerging standards such as SAML (Security Assertion
Markup Language). Other omissions include workflow, business
process management, and synchronous and asynchronous transactions.
Alongside Java and .net, these are - or should be - critically
important to users looking to start building and deploying Web
services.
So, given that WS-I is so important, what must it do to thrive?
First, it must avoid appearing to dictate and direct standards
initiatives, while also playing the role of diplomat and arbiter
when dealing with overlapping standards under the control of
different standards bodies. This is particularly important and
tricky because WS-I is dependent on the responsiveness of the
standards bodies. The World- Wide Web Consortium is yet to
establish a Web Service Definition Language working group - a year
after the publication of the original specification.
WS-I must also quickly win consensus among traditional competitors
with vested interests in driving the standards to meet their own
needs. Interoperability will increase the size of the Web services
pie and will help to drive co-operation but WS-I must not
underestimate the politics involved.
Who's who in WS-I?
Suppliers:
Accenture, BEA Systems, Fujitsu, HP, IBM, Intel, Microsoft, Oracle,
SAP, Autodesk, Cape Clear, CommerceOne, Compaq, Groove Networks,
Jamcracker, Plumtree, and Verisign
Users:
Daimler Chrysler, Ford Motor Company, Reed
Elsevier, Reuters and United Airlines
Neil Ward-Dutton is research director at Ovum