

Thin-Client Computers Come of Age
May 3, 1999
|
Additional Information
|
|
Side Bars
|
Windows NT Terminal Server Edition and Citrix MetaFrame: Competitive or Complementary?
|
|
PDF Chart
|
MetaFrame extended features
|
|
Related Links
|
Network Computing, Nov. 1, 1998
"Windows-Based Terminals: Construction in Progress"
Network Computing, Oct, 15, 1998
"Java Brews Up a Storm in the Enterprise"
Network Computing, July 15, 1998
"Desktop Management: Squeezing the PC, Not Your Users"
InternetWeek, June 15, 1998
"Thin Clients: Desktop Users Slim Down"
Information Week, Oct. 12, 1998
"A Thin Foothold"
|
By Dave MoltaWhen Apple Computer's Newton PDA (personal digital assistant) was first released, the hype was intense. But despite all the enthusiasm, the product had little impact on the market. Many pundits wrote off the PDA as a bad idea whose time would never come. Today, the success of 3Com's PalmPilot, and, to a lesser extent, similar devices running Windows CE, demonstrates that the market need was there all along, but the products weren't quite ready.
A similar phenomenon occurred with thin-client computers. Hyped extensively before their time by major vendors, including IBM Corp., Oracle Corp. and Sun Microsystems, the initial entries into the market were greeted with skepticism or disdain. They were slow, complex, feature-limited and not nearly as easy to manage as the market had been led to believe. Meanwhile, prices for traditional PCs were in freefall, and the old rule of thumb that historically guided managers to budget $2,500 for new PCs was being challenged by fully configured systems selling for less than $1,500. Equally significant, the lack of a major upgrade to the Windows OS during the past three to four years was allowing managers to stretch the life cycle of these machines. All of a sudden, budgeting for PC replacement every three to four years didn't seem quite so silly anymore. Given annual amortized costs of less than $500 per machine, why would anyone want to give up full-function PCs for thin clients?
The answer to that question has two parts. First, though it is true that PCs are less expensive than they've ever been, they are not significantly easier to manage than they were when Windows95 and Windows NT were introduced. Despite efforts to throw software at the problem in the form of desktop-oriented management systems, many IT managers continue to struggle in their efforts to deliver up-to-date and reliable software services to desktop users. At the other end of the Ethernet, many users are just plain frustrated with system instability, particularly those who view the PC simply as a tool to get the job done. A well-managed thin-client platform is often regarded as an ideal solution for these users. It's reliable, secure and predictable.
In addition to existing general-purpose PC users, there is another prospective audience for thin clients--the millions of users who still interact with dumb terminals attached to host computer systems. While easy to maintain, dumb terminals impose obvious limits on the kinds of applications that can be deployed to serve user needs. The prospect of replacing these devices with PCs--especially in remote and branch-office locations where support isn't readily available--is not very appealing to IT managers.
TCO Under a Microscope Analysts have had a field day exploring the many dimensions that characterize TCO (total cost of ownership) of computer systems. TCO is really just a combination of capital costs for computer equipment, support costs to keep that equipment running and lost productivity accruing from system failures. Thin clients are not a magic pill when it comes to reducing TCO. In fact, the costs for thin-client desktop systems are not significantly less than for value-priced PCs. Because these systems often require additional server capacity, savings at the desktop can be quickly offset by higher back-end costs. And despite a recent price restructuring on Windows NT TSE (Terminal Server Edition) that makes it significantly more affordable, software licensing costs can be equal to or greater than the cost of software licensing for traditional PCs.
The benefit lies in control. Much like the host systems of the past that established standards for reliability, thin clients are often regarded as being capable of providing consistently high levels of service, mainly attributable to centralized management. But just as with minicomputers and mainframes, this reliability requires expertise in managing a new hardware and software environment. Today, that expertise is often lacking inside organizations, which sometimes need to look for support from the many consultants, integrators and VARs (value-added resellers) that are quickly capitalizing on this growth market. In the end, reliability will be a product less of the technology and more of the people who implement and support it.
Despite its positive potential, an agenda of replacing PCs with thin clients in order to minimize TCO is often a difficult sell politically within organizations because it diminishes desktop independence. Thus, thin-client initiatives are often most successful in organizations that have clearly defined management hierarchies with strong executive decision-making.
|