Users stand to win cost cuts and efficiency gains from Web
services, but only if they can manage the complex software
relationships effectively
Web services have been around since the mid 1980s when people
started to use electronic data interchanges to facilitate
transactions. The emergence of the Internet and HTTP as the
ubiquitous communications protocol has rekindled the promise of
truly open communication and interoperability between applications.
This time the promise is of a new magnitude. An ultimate scenario
is being painted of companies and consumers seeking applications or
services on the fly that they can use to perform a variety of
transactions. The behind-the-scenes complexity reaches new heights
with enterprises dynamically seeking out multiple Web services
components, which would work together to form one giant application
- managing commission payments for the salesforce, for example.
Interestingly, recent research conducted by business applications
specialist Compuware reveals it is small and medium-sized companies
that cherish the highest hopes for Web services with 53% viewing
increased revenues as a key business benefit of the technology
compared with 40% of large organisations. Improved collaboration is
the main pull for 68% of large organisations while 64% expect Web
services to reduce business costs.
"This research suggests that it is the smaller, more agile
companies that are most likely to view Web services as an
opportunity to not only generate revenue, but increase their
business responsiveness and collaborate with partners and
customers," says Mike Lucas, chief executive of Compuware. However
seasoned industry watchers say this is because smaller companies
have no real experience of the realities of integration.
Watch out for obstacles
Indeed before companies, large
or small, can reap any business benefits from Web services, there
are a number of obstacles to overcome. Managing multiple pieces of
software on a day-to-day basis is a whole new ballgame for which
there is no blueprint according to ZapThink, analysts specialising
in Web services and XML. "Implementation is still a custom
exercise. There is no existing best practice in this realm," says
principal consultant Jason Bloomberg.
Danny Goodall, EMEA marketing director of Web business specialist
Sonic Software, points out that HTTP is a very low-level standard
indeed and in the hierarchy of communication, is roughly equivalent
to making sounds. "Higher up the stack are words, and above that is
language." In other words the hard work of translating and
transforming these HTTP- wrapped objects into something meaningful
still has to be done.
For example, should a purchase order arrive at a company, the
stock-checking application on the mainframe has to understand it
and make the appropriate calls. Only at that point may a credit
check be necessary and depending on the outcome, the order may be
routed through to accounts. All these applications are likely to be
written in different languages and that all adds up to a lot of
translation.
The great unknown
Systems integrators are the most
fearful of the brave new world of Web services as they would bear
the brunt of exposing all legacy systems to HTTP. Chris Read, head
of digital solutions at Andersen, queries exactly what all this
work would be in aid of. "What is the killer app going to be for
Web services? I haven't found anyone who knows. The fact that you
can knock up a piece of software that can talk to anything is the
start of the problem," he contends.
Read believes that while all the excitement has been around the
discovery and ratification of an enabling technology it is doomed
to failure because business processes are unclear. "No two
companies want to do things in the same way," he says.
"The problem with Web services is that it assumes that if, for
example, you want to buy tin from Venezuela, you just launch a Web
agent and plug in," says Read. Real life isn't like that because
you can't trust a supplier's fulfilment capability. There are a lot
of contractual and legal obligations that Web services just do not
answer," he says.
ZapThink agrees that the business processes have not been thought
about sufficiently and singles out the issue of payment as most
lethal. "If you are aggregating Web services from different
suppliers for your application, and forget to pay for one, then the
whole application might fail," Ron Schmelzer, a senior analyst
says.
Schmelzer points out that software metering never worked for
application service providers and consumers don't want to use a
credit card to pay for a product worth under a pound because
they'll lose money. "If you're going to rent a sales automation
service for one hour, you're going to have to set up an agreement
with the supplier first, and agree any discounts at that point," he
suggests.
Behind closed doors
While all the thorny business
issues are being thought about and discussed, for the time being,
large corporates are deploying Web services as an internal
exercise. This phase in Web services represents an opportunity to
get a handle on technical aspects like network traffic, managing
loads and security.
The new messaging standard, Soap, doesn't make a great deal of
difference to network managers, says ZapThink's Schmelzer, as it
merely replaces Dcom or Corba. But the addition of XML, a clunky,
text-based format designed to be read by humans as well as
computers - and with WSDL on top - could potentially clog up the
network and will need careful managing if traffic loads get heavy.
However, the biggest management headache of this new model of IT
may prove to be loss of control over software functionality. The
focus in-house would shift from software development to the testing
and assembly of various components. It could be the ultimate game
of detective work when a bug occurs and there are 30 components
involved. For Read, it is the ultimate nightmare: "What do you do
in the middle of the night when your key application fails? There's
no source code and no fix."
With this in mind it seems we are experiencing the lull before the
storm and whatever obstacles exist, businesses will increasingly
demand Web services. As Schmelzer says: "Users want to save and
make money today - and Web services has a story - it's not just pie
in the sky."