Java, An Introduction

Last month we compared groupware with intranets and touched on the development of Sun Microsystem's Java. This time around I'd like to delve a little deeper into the subject with a discussion on Java development and deployment and it's effect on the enterprise. Those who believe Java to be a method of spicing up your web site with some animation and other such window dressing have completely missed the point. Certainly Java can accomplish this (there are also easier and faster methods) but it's developer's vision extends far beyond this modest goal. Officially released May 23rd,1995, Sun's Java has matured from an interesting Internet rooted curiosity to an enterprise computing platform being deployed in some manner by over 67% of Fortune 1000 companies... all before it's second birthday.

What is it anyway?

Simply put, Java is a robust programming language similar to C++ in that it is object oriented and uses like syntax but that is where the resemblance ends. The key difference from all other programming languages is that it is platform independent. Let that sink in for a moment... this one feature alone has far reaching implications. Software developers need not port their product to a variety of computer platforms in order to secure market share... one version will work on UNIX based servers, PC's running any OS, Macintoshes, dumb terminals...whatever. This permits software engineers to significantly reduce product development cycles and enables them to concentrate on quality vs. quantity of stations able to run the software they are producing. Companies embracing this technology are varied indeed: manufacturers of computer software, PDA's, cellular phones, smart appliances, telecommunications hardware... you name it.  This is only the beginning.

Cost of  Ownership

As the evolution of corporate networks has hampered legacy and disparate system's integration into the networked pool, Java's cross platform functionality finally promises true heterogeneous interoperability. This advantage will free network administrators from the daily headaches involved in maintaining legacy systems vs. the current state of the art. One set of core applications may be implemented across the enterprise; reducing cost of ownership through the decreased knowledge base required to maintain and support them.

A number of  commissions investigating life-cycle costing of a typical node on a LAN (including the best known Gartner Group study) discovered the sum to be between 10 and $15,000 per station per year (Gartner estimated $11,900). This expense is primarily comprised of service calls to the corporate "Help Desk": due to the fact that the average user (typically not well familiar with the inner workings of his/her machine) has nearly complete control over the nature and configuration of the software residing on their station. This autonomy is also a great network security risk; as people bring in disks from their homes, potentially carrying viruses that could bring down the entire system; resulting in costly downtime.

With Java and a new stripped down computing system called a "thin client" or network PC, system administrators regain the control they once had in he early days of  the client server model. There are no user input devices other than the keyboard and all the software selection is done by the administrator. The thin client requires little inherent processing power; as in the "Java Computing" model, all processing of applications, file and print services, program storage and data warehousing is done at the server level. This results in the centralized administration of all facets of the information technology infrastructure by those who are best able to perform the task. Software glitches, file and print service snafu's, and network downtime are thus reduced to a bare minimum.

An independent study commissioned by Sun Microsystems, estimated the cost of ownership of the typical Java Computing model of enterprise IT deployment (employing thin clients) to be $2500 per station per year... a dramatic difference to say the least.

What about Active-X?

A common misconception is that Microsoft's Active-X technology is in direct competition with Sun's Java. The basis of this misconception is the fact that on the web, Active-X and Java can perform some of the same functions yet while Java is a complete programming language, Active-X is a component framework which enables applets written in a variety of languages to become active web content. This means that Active-X components may even be written in Java. Microsoft has invested a great deal of energy in speeding the development of Java with the introduction of it's own superset of the language called J++. This version contains many of the Windows specific objects that Sun (surprise, surprise) neglected to include.

A few caveats

While Java is indeed being developed and adopted at a breakneck pace; true large scale applications are scarce. Part of the reason for this is the fact that due to it's nature, Java is many times slower than programs precompiled under the current standard, C++. Microsoft again is trying to address this issue with a new "Just in Time" (JIT) compiler which speeds operations considerably. Sun has taken a different tack; recently announcing the release of inexpensive "Java Chips". Sun promises these IC's will eventually speed the operation of Java to a level comparable to C++ and beyond. While Sun claims that these chips will soon be commonplace in standard computer motherboards, it clearly supports the thin client model.

With JIT, Microsoft currently runs Java at speeds close to the first generation of the chips and industry analysts are loath to predict which technology will eventually become the standard. Contrary to Sun's assertions that Java is being rapidly adopted as the core framework for large scale enterprise IT solution deployment, most CIO's are sitting back a while to see what lies around the corner for this emerging technology. If, however, the Java Computing model lives up to it's promise of significantly reduced cost of ownership, look for a thin client to appear on your desk in the near future!

Originally published in Monitor magazine's lanStuff column, December, 1996, by technology columnist, Ray Richards.

• Top
• Navigation
• Accessibility
• Colophon