
Corporate.Net
Are NCs Really Worth The Price Of
Admission?
But be aware that things aren't what they seem to be. There's no law that says a company can't take an old technology and re-release it as new. Is it an X terminal or is it a network computer? And if a standards-based desktop NC is what you're after--a pure Java computer--don't lose sight of the fact that the applications you need are probably still a twinkle in developers' eyes.
Total Cost of Ownership
The fuel that is fanning the flames of the NC fire is desktop computer total cost of ownership. The TCO issue has been articulately stated in analyses performed by Gartner Group and other research and consulting firms, suggesting that the average networked PC costs $10,000 to $15,000 annually for hardware, software, administration, support and user operations. Of this total, studies claim, a relatively small percentage of the TCO is attributable to hardware and software costs. The brunt of the
se expenses is attributable to administration and support. Although there has been significant debate over the numbers, experienced IT managers have long realized that capital costs for PC hardware and software are very much like the tip of an iceberg. Look below the water line to get a view of the real money you'll spend in network and system administration and lost worker productivity as users strive to optimize their system environment. The reality is we have moved from a mainframe environment where we had a staff of systems administrators whose numbers were often in the single digits for a midsize organization to the point where we have thousands of systems administrators: one at every desktop.
Early proponents of the NC, including Oracle's Larry Ellison, garnered attention in the media with promises of a $500 dynamo, but this is at odds with the TCO premise. Never mind the fact that Ellison and friends may have left out the cost of the monitor and adequate memory to run more than a single applicatio
n in their initial cost estimate. That $500 number rang a chord in the minds of many
managers who signed purchase orders for those $2,500 Pentium PCs. To this day, people still seem overly fixated with those above-the-water-line costs. However, if NCs are to become more than a blip on the distributed computing map, the payoff must come in reduced operating expenses. That's where the real money is being spent in most organizations.
Despite the potential savings, it may take some heavy-handed decision-making to turn the promise into reality as many users are sure to resist a CIO's attempt to take away the power and flexibility of today's desktop computers. And there is often a hidden agenda among many knowledge workers: a desire to become a computer guru for their departments, a responsibility that often carries with it a fatter paycheck and greater opportunities for professional advancement. Although these users are clearly a minority, their influence at the departmental level cannot be ignored.
Talk t
o Intel and--surprise, surprise--it will tell you that there is a relatively small market for limited-function NCs, perhaps 10 percent to 15 percent of desktops. Intel views these devices primarily as replacements for dumb terminals and possibly desktop computers that have traditionally delivered a very limited set of applications. Even factoring in their obvious lack of objectivity, they may have a valid point. At the same time, they probably overlook the potential benefits of such devices in organizations that are geographically dispersed. A branch office with 10 or 15 employees often is too small to justify the administrative costs of installing a traditional network server and may lack the expertise to maintain desktop computers providing access to a range of applications. Ship a dozen boxes filled with NCs, plug them in, connect them to a corporate WAN or the Internet via an ISDN-based router, and voila!, instant computing with minimal muss and fuss.
The Core Issues
So what are people lookin
g for in an NC? Everyone wants to find a way to reduce cost of ownership, but what is the
magic feature set that spells winner in this market? Many vendors are wrestling with this question. Do you make a pure play based on Java or take a more conservative approach that provides better access to legacy applications? Do you fight the Microsoft applications juggernaut or do you find a way to peacefully coexist? Once you decide on the NC's capabilities, the implementation issues become paramount. You may not be able to afford to add functionality if it comes at the expense of administrative complexity.
Most vendors are taking a pragmatic approach and realize that there is a fertile market for replacing the millions of dumb terminals that still populate desktops. As such, providing terminal-emulation capabilities is critical. Expect to see VT-100-style emulation to provide access to host-based applications on Unix systems as well as TN3270 emulation for access to IBM mainframes.
To download an Adobe Acrobat .pdf format version of the Network Computer features chart, click here.
Network Computer Back-End Application Services
|