The term “Web Services” has begun to take on a familiar ring, becoming
accepted in the business landscape as jargon to mean anything and everything
delivered over the World Wide Web. But apart from their generic marketing
use, the words have evolved into a more specific definition.
So what exactly is a web service? Quite simply, it is a fundamentally new
approach of developing a software application that can share information
through Internet Protocol (IP). What makes all of this so revolutionary is
that these newly created systems would be able to interact and exchange
information regardless of the platform or environment.
In doing so, the Internet would evolve from being simply a place that you log onto in order to retrieve information. “We will begin to think of the Internet as a communications infrastructure that allows us to communicate,” said John Patrick, the former vice president of Internet technology at IBM who retired after a 35-year tenure. He is the author of Net Attitude.
The one commonly cited example of web services to illustrate the innovation is the
calculator that has been pre-programmed to know every jurisdiction’s tax
code in the U.S. Assuming this calculator was placed on the web and programmers knew how to find and remotely access it, a company could simply tap into this fictitious calculator
for a nominal fee with their Great Plains accounting software. In doing so, the company could figure out exactly what net income and cash flows were at any given time.
Now, you might think: “Big deal?!? My Quicken 99 has been able to
download data from the Internet for years!”
But then you’d be missing the point. If that communication could occur no
matter what the operating system or regardless of what language the
application is written in, that would simplify systems integration and
seamlessly open new doors for business and even consumers.
At the heart of the innovations lies a new standard that has been in
development for years but painstakingly slow to actually deploy called
Extensible Markup Language (XML). As opposed to Hypertext Markup Language
(HTML), which describes what a web document looks like, XML tags data in the
proper context to tell you what that document means. With the backing of the
entire industry, XML creates advantages that are so obvious that even its
biggest critics can understand the operating efficiencies.
Yet, corporate leaders and IT managers aren’t completely convinced that
web services are the next major IT revolution on the horizon. Analysts like
Forrester Research and Gartner believe deployment will continue at a
snail’s pace this year. As proof, despite massive marketing efforts by IT giants, a
staggering 75 percent of InternetNews.com readers (which constituted
primarily corporate managers and IT executives) said they are either not
interested or just looking now at implementing web services, according to a
recent, informal poll.
That has led some people to argue the promise of web services have been
over-touted by the software development kit (SDK) vendors as well as by the
general media and investors. IT executives and corporate managers have been
slow to embrace the concept fearing a reprise of the dot-com bust especially
in light of economic uncertainty. But while very few individuals, if any,
have started a business selling commercially developed web services on call
for corporate portals, companies like NetEdge Software of Wake Forest,
N.C., have flourished helping companies use web services internally to integrate their clients’ legacy systems.
“We are pitching web services as an enterprise play. The goal is not to
build services off of eBay or something like that but build interoperability
behind the firewall,” said Jay Pitzer, vice president of sales and marketing
at NetEdge. For example, the company has used Web services to help a major
pharmaceutical giant integrate results from clinical trials.
“Interoperability is very important because in clinical trials, data is
coming from different systems. Web services is a beautiful play because you
can build a web service using WebSphere and as long as they are WSDL
compliant, it should work where ever you are,” Pitzer noted.
How does XML speed integration? And what can it NOT do? See Page 2.
What’s behind XML?
XML web services are definitely important but if you were to read the
vast amount of material about the subject found on the Internet, you would
think that it is all a developer would need to know. Far from it! Every
developer creates software application by building a main program that
houses your primary business logic (a.k.a. project mission) and adding
various components such as database integration or a graphic user interface
(GUI) to create screens that the end-user would see.
For example:
main()
{
Initialize_Params(); <----setup everything (variables, etc...) Draw_GUI_Panels(); <----allocate video memory for API objects GUI_Loop(); <----Loop and wait for key-click or mouse-click }
But while XML does allow different platforms to transfer data back and forth, it still doesn’t allow one application to communicate with another on all levels. In developer-ese, it means you still don’t have “native access” between applications. Developers still need to have their XML data packaged and parsed as well as distributed in a native platform. But it is this particular nuance (how XML data is packaged versus how it is
distributed) that is responsible for much of the confusion in the burgeoning
web services field, according to analysts and industry officials.
“XML data needs to be packaged and parsed. This process does impose
overhead. [XML] is a standard that the developer will use; however,
that doesn’t mean performance will be optimal if the developer requires
tighter integration,” said Don Hsi, president of Halcyon Software, which
makes developer tools to help programmers migrate from .NET to the Java
platform.
But how exactly the data is transferred or distributed is the real beauty
of the XML standard and why it forms the foundation of Web services. The
World Wide Web Consortium is expected to formally recommend a communications
protocol called Simple Object Application Protocol (SOAP). Thanks to SOAP,
different systems can share information regardless if the application is
created in Java or mainframe COBOL. But SOAP alone isn’t enough to
revolutionize software development because the XML data still needs to be
encoded and parsed into the client application. To do this, information is
spelled out in the Web Services Description Language (WSDL) document, which
describes the web service’s interfaces in enough detail for the application
to talk to it. The last common element that web services are designed to
share is the Universal Discovery Description and Integration (UDDI) –
literally, a directory of web services that have become a marketplace all
its own.
To be sure, for the XML data to be useful to anyone it still needs to be
interpreted using applications or systems created out of Visual Basic, Java
or COBOL.
“A web service is about interoperability…not portability,” said David Truog, analyst at Forrester.
For this reason, the software vendors – ironically, those same companies
originally responsible for the standards – like Microsoft, Sun Microsystems
or IBM have had their marketing machines in overdrive touting tools in hopes
of capturing the support of the various developer communities.
Unfortunately, the negative side effects of all the marketing are
allegations of hype or vaporware and confusion about how .NET is supposed to
interoperate with Java.
Which platform is right for your organization? See Page 3.
The Mother of All Web Services
So which web services developer platform is right for your organization?
To answer that question, a little developer history provides helpful
context. When Java first came out in 1996, applications developed using the
language resided on the client side and were considered to be much too slow
to be viable solutions by serious developers and IT managers. The principle
reason was because of the GUI (pronounced Goo-ie). However, as the industry
migrated toward server-side computing and Java went deeper and deeper into
an enterprise, its utility increased because machine-to-machine
communications don’t rely on visual components. Therefore, no GUI was there
to slow down a request processing.
With the concept of “write once, deploy anywhere,” all of the sudden
companies like BEA Systems (WebLogic), IBM (Websphere) and of course Sun
were able to capitalize on the Java-based applications servers they had been developing.
When Microsoft unveiled its .NET strategy to transform their software into
online service, Java partners banded together and turned Java 2 Enterprise
Edition (J2EE) into a rallying cry. So much so that companies like Rational
Software have placed major bets that .NET and J2EE will be the last men
standing when the web services war is over. However, .NET and J2EE are, by
far, not the only two frameworks available, as illustrated by SYS-CON Media’s Web
Services Journal.
“Leadership will come from vendors other than Sun through 2003,” said David Smith, vice president and research director for Gartner.
Next week, Microsoft plans to unveil Visual Studio.NET, the mother of all
developer tools for the mother of all platforms. With VS.NET, Microsoft is hoping to convert 7 million licensed developers to upgrade to the .NET platform. In fact, Microsoft is in the unique position of having the ONLY platform that is capable of directly reaching PC consumers. But, until now, most of the world’s largest developer community haven’t yet experimented with web services. There are, however, armies of developers that are solely dedicated to building one component of a software app after another — components that can easily be turned into web services.
For the time being, individuals interviewed for this report declined to
predict the future viability of nascent web services segment. Most research firms predict 2002 will be a year to learn, not implement. Jupiter Media Metrix believes web services will not begin to affect companies’ broader business decisions for another 18-24 months. Gartner thinks web services will sooner or later dominate new application deployment but not until 2004.
In conclusion, it will likely take more time for businesses to sort the marketing hype to find the real technological benefits. Web services will mostly trickle rather than rush
into the corporate IT department but it has already proved it won’t go away any time soon.
— Prepared by Robert Liu, Steve Kapsinow and Dan Muse with contributions from Thor Olavsrud and Jim Wagner. This is the first part of a 3-part series. Next week, we will examine how web services are evolving and look at what links are still missing to accelerate deployment. In addition, we will examine the unique role that Microsoft plans at this burgeoning segment of the software industry.
