The Attraction Factor
The X and ICA paradigms are appealing as a
desktop solution not for their technical, wire-level technology, but for their
ability to deliver NT-based applications to a variety of desktop equipment.
Without X or ICA, desktop access to NT-based applications is limited to desktop
systems that run either NT Workstation or Server or that use a remote-control
product such as Symantec's pcANYWHERE or Avalan's Remotely Possible. (For more
on these products, see John Enck, "Symantec's pcANYWHERE32," and "Avalan's
Remotely Possible/32," May 1996.) The problem with the first option, of
course, is the cost of configuring a desktop system so it can run NT.
The problem with the second approach is that remote-control products are
not for high availability, high performance, multiuser access (however, they are
good at letting a single non-NT desktop access an NT application). In contrast,
with X or ICA, you can access NT-based applications with X-terminals; Windows
(ICA) terminals, such as the ones Wyse and Tektronix manufacture; UNIX systems
running ICA client software or X-terminal emulation software; Macintosh systems
running ICA client software or X-terminal emulation software; Windows 3.11 or
Windows for Workgroups (WFW) systems running ICA client software or X-terminal
emulation software; or Win95 or NT systems running ICA client software or
X-terminal emulation software.
If you choose a server-side product that supports both ICA and X technology
(e.g., Insignia's NTRIGUE and NCD's WinCenter), you can mix and match desktop
strategies. You can deploy X-terminals, Windows (ICA) terminals, ICA client
software, and X-terminal emulation software as you need. In contrast, if you
choose a product that supports only ICA (e.g., Tektronix's WinDD, Wyse's
WinFrame, or Citrix's WinFrame), you'll be locked into one desktop paradigm.
Note that most WinFrame-based software vendors ship DOS, 16-bit Windows, and
32-bit Windows ICA clients with their server software. These vendors typically
provide client software for other environments (e.g., Mac or UNIX) as an
additional, priced item.
In theory, X/ICA access to NT is a good idea. To test the theory, Windows
NT Magazine brought the following products into our Lab for evaluation: the
Wyse 2500T ICA terminal and Wyse's Version 1.5 implementation of WinFrame (for a
case study of a Wyse and WinFrame solution, see Kevin Woodward, "ProStaff's
WinFrame Solution," page 148); NCD's Explora X-terminal, WinCenter Pro 1.6
(NCD's version of WinFrame, which includes X support), and NCDware (NCD software
that lets X-terminals boot from an NT server); a Tektronix Netstation 400 Series
X/ICA terminal, which supports both paradigms, and WinDD 2.0 (Tektronix's
implementation of WinFrame); and Insignia's NTRIGUE 2.0 (Insignia's version of
WinFrame with X support).
These products' ability to deliver NT applications to non-NT desktops
im-pressed me. (For information on the specific products we tested, please refer
to the product sidebars.) We used a variety of X-terminal emulation software,
including Insignia's NTRIGUE Mac Client, NCD's PC-Xware, and Hummingbird's
Exceed to test X-to-NT access. We tested only interoperability among the
server-side products and therefore did not include them in the product sidebars.
Is X/ICA Cost Effective?
Are you really saving money if you
adopt an X or ICA desktop strategy? Well, the answer depends on where you're
starting from and how you look at cost savings. If, for example, your
organization has X-terminal systems, 386 or 486 PCs, or Macintoshes on the
desktop, the X and ICA strategies let you introduce NT applications without
spending money to upgrade your desktop hardware. In most medium- to large-scale
organizations, this approach will be more cost effective than placing NT on
every desktop--even after you figure in the higher cost for the software and
hardware you need to drive the X or ICA server.
But what if you're building a new solution? Is the X or ICA approach always
more cost effective? If you look at the equipment costs only, the savings
depends on the scale. You can purchase a Windows (ICA) terminal for less than
$1000, which will clearly help offset the cost of a high-end server. If you
broaden your vision of cost to include the price of supporting and maintaining
desktop equipment, X and ICA terminals reduce these costs because they have no
local hard disk to maintain (or corrupt), no floppy drives to introduce games or
viruses, and fewer moving parts to break.
Would you ever deploy X or ICA on new, full-scale PCs on the desktop?
Probably not, but you might still find some useful applications for X/ICA on a
smaller scale. For example, even if you roll out new PC hardware for the
desktops, certain key applications may restrict you to running Windows or Win95
on certain desktops, or you may need to accommodate remote access from traveling
laptops or telecommuters. X or ICA access is a good option in these situations.
One emerging application for ICA (but not X) is to have an ICA client
plugin for your Web browser deliver applications over the Internet. When you
click an ICA tag in a Web page, your browser launches your ICA client helper
application, and bingo--you're running an NT application over the Internet. The
possibilities for this technology range from software vendors offering product
demos, to corporations offering employees and partner companies full-blown,
multiplatform applications (remember, ICA clients are available for many
hardware platforms). Of course, running ICA over the Internet comes close to the
third paradigm, Web browsers. But, as you will see in a later article, the Web
browser paradigm has a life of its own.
So when high noon comes in your organization, will you be facing this gang
of desktop paradigms? Chances are pretty good that you will. And considering
that you've probably already faced the Internet connection challenge, the
TCP/IP-only network challenge, and the NT vs. NetWare challenge, chances are
pretty good that you'll come out a winner again. Just remember to keep the sun
at your back.