Web Services: A Reality Check
Introduction
A popular vision of cyberspace is that
Internet content will become a utility provided like electricity or water; you turn
on the tap or flick the switch and what you need is delivered.
Unfortunately, with the Internet, it's not that straightforward. Electricity and
water were provided by monopolies that defined a standard and implemented it. Everyone
got the same thing and generally it worked. What's more, the complexity was manageable
because only one service was being provided.
Efforts have been made to introduce standards to the Internet, but with only limited
success. You need only compare how some web pages are oddly rendered in different
browsers to see how difficult it is to maintain standards globally.
And content display is just the tip of the iceberg. Things start to get really complex
when we decide to share information between various applications and devices, or
we want to streamline the process of searching for, filtering and channelling information.
As with display, efforts at standardisation have been made. For instance, XML has
set standards for describing the content and structure of data, and both XML and
RSS (Really Simple Syndication) have made possible a growing acceptance of data
feeds. This is where information is provided to consumers rather than having to
be searched for and collected. However, data feeds are limited because the information
received by consumers is determined by the provider.
By and large, there is still too much human involvement in processes that could
be entirely automated. In other words, the Internet gives us access to a phenomenal
volume of content, but it's a long way from being optimised.
The promise of interoperability
Tim Berners-Lee, considered to be one of the founders of the Internet, has the vision
of a 'semantic web' where meaning is attached to data and data is interoperable.
One example would be how a calendar program interacts with data. The calendar software
may have no idea what a series of digital photographs are or what a bank statement
is but it does recognize that both have dates and therefore can use the information.
Suddenly your calendar becomes a diary, a record of the pictures you took on a particular
day and a virtual bank statement. In the future, data will be better defined and
software will be able to make more intelligent use of it.
One development that may take us closer to data interoperability – by capitalising
on the XML standards – is web services. IBM's description of a web service
is a 'self-contained, self-describing, modular application that can be published,
located, and invoked across the web'. The application may be a simple service or
a complex business process.
The implications are huge and idealists embark on flights of fancy. Indexed services
will become globally available for use by anyone to perform a function. There will
be no need to reinvent the wheel. Any service you require will be available on the
Internet and everything will be seamlessly integrated.
The obstacle course to interoperability
But there are obvious problems which dampen the idealist's dream. Will the service
be available when you require it? Can you trust the provider? How much will it cost?
Microsoft's Passport is a good example of users' concerns. Theoretically you can
now delegate authentication to Passport. It is a free service and most likely will
be available when required, but it comes as no surprise that many are reluctant
to trust another organization with this particular responsibility.
A less idealistic vision of web services is that they will mainly be used within
a business or for linking to trusted external clients. They will provide the mechanism
for utilising data from disparate systems. Data defined to a standard can be freely
used without further manipulation.
For example, a business need no longer hold exchange rate information. When the
data is required, a request is sent out to an exchange rate web service and the
necessary calculation made. Theoretically this should all be automatic – the
required service is found in an index, the descriptor explains which inputs to provide
and what outputs to expect, complete with the billing information.
In the short term, however, these services will generally only be invoked with prior
agreement and discussion between the involved parties. Few are ready for software
that searches out the best deal and selects it with no human intervention. Imagine
the concerns over receiving an invoice that is automatically imported into your
payments system, from a provider you know nothing about for a service you weren't
even sure you requested.
Even so, in a controlled environment, complicated processes between businesses can
be automated and implemented with little IT development effort, if your data and
interfaces conform to defined standards.
Conclusion
Web services promise greater global integration with less replication of data and
processes. As with any new technology, the benefits are directly proportional to
the degree of industry acceptance. If web services become the norm, as many believe
they will, data exchange will be greatly improved and there will be far more productive
use of information with greatly reduced human involvement.
|