Digital Intelligence & Forensics
Measurement & Analysis
Performance & Dependability
Process & Performance Improvement
Security & Survivability
NEWS AT SEI
This article was originally published in News at SEI on: June 1, 1998
“In the future, network computers will be purchased
and used with the same enthusiasm as home
Scott Adams, The Dilbert Future: Thriving on Stupidity in
the 21st Century (New York, NY: HarperBusiness, 1997).
Is Mr. Adams right? Maybe. But it’s important to
make a distinction between network computers
(NC) and Net-centric computing (NCC). An NC is
just one type of “thin client.” NCC is more than just thin clients; it is an
emerging phenomenon whose effects will be profound and far reaching.
Almost anyone involved in computer science, information technology
(IT), or software engineering will be affected.
So what is NCC? The underlying principle behind NCC is a distributed
environment where applications and data are downloaded from network
servers as needed. This is in stark contrast to the current use of powerful
personal computers (PCs) that rely primarily on local resources. In some
respects, NCC resembles an earlier computing era of mainframes and
dumb terminals. However, there are important differences. NCC relies on
portable applications that run on multiple architectures (“write once, run
anywhere”), high bandwidth (for downloading applications on demand),
and low-cost thin clients such as the NC, the NetPC, and Windows-based
SEI Interactive, 6/98
An NC uses Java for local processing. It was initially proposed by the
“gang of four” (IBM, Netscape, Oracle, and Sun) as an alternative to the
Microsoft/Intel duopoly. The NC vision was for a computer that did not
run Microsoft Windows software, and it could use processors other than
Intel's Pentium chips. This vision has at least one flaw: People still want
to access their legacy applications (primarily Windows programs) and
The NetPC was an interim solution proposed by Microsoft and Intel to
counter the NC. It is essentially a stripped-down PC with a sealed case.
The selling point of the NetPC is a reduction in total cost of ownership
because end users are unable to add or remove new hardware or software.
It can be centrally administered and it can run Java applications if
needed. In this sense, some consider it to be a better network computer
than the NC itself.
The thin client that seems most likely to succeed is the WBT. A WBT is
the thinnest of the thin clients, relying completely on a central server
for applications and data. The WBT acts only as a display device, much
like an X Station does on Unix. Early versions of WBTs are in fact
current computers running Citrix Systems' WinFrame client. This
application lets users on Windows, Macs, or Unix machines access a
centralized computer running a modified multi-user version of Microsoft
Windows NT Server. This technology (code-named “Hydra”) is being
rolled back into Windows NT 5 Server as the “Windows Terminal
Server.” I think it will prove successful because it lets organizations
leverage their current IT infrastructure. The productive life of older 386-
and 486-based PCs can be extended while newer machines are
Effects in the office
Irrespective of which type of thin client you adopt, how will NCC affect
you in the office? Since almost everyone is affected by software these
days (like it or not!), you likely fit into at least one of the following three
categories: user, developer, or administrator. For some users, the relief at
not having to maintain a PC means they can instead concentrate on their
primary tasks. Other users may chafe at the limitations that NCC brings
with it. The removal of “personal” from PC means that users will no
longer be able to significantly alter their desktop environments.
For developers, NCC offers an opportunity to greatly increase their
customer base: An application written in an NCC-aware programming
SEI Interactive, 6/98
language, such as Java, means writing code once and having it
immediately accessible on multiple platforms. It also means a different
development environment, a new deployment model (renting
applications versus buying), and new concerns about security. In a Netcentric
world, security is not just for system administrators. For most
developers, security is a quality attribute that is treated as an add-on to
the system. For NCC, it needs to be treated as a first-class concern.
For administrators, NCC means a potential reduction in the cost and
complexity of managing IT resources. The total cost of ownership issue
has been cited as one of the motivating factors behind NCC, but so far
little real data is available to suggest that NCC will be cheaper than
today’s methods. It may in fact be more expensive because of the
increased complexity of a heterogeneous and distributed environment.
Outside of the office, the effects of NCC may prove even more
significant. For software engineering, NCC offers a fundamentally new
way of thinking about software. Basic issues such as version control need
to be re-evaluated. For example, if a software application is being
delivered to the user (and continually updated) using push technology
such as Microsoft’s CDF or Marimba’s Castanet, what does it mean to say
“the current version”? If the application is being monitored and updated
in the manner of superdistribution, this may make software versions
based on millisecond differences a reality.
Still not convinced this “NCC thing” will really affect you? Here’s
another quote from Mr. Adams's book:
On the off chance that you are not familiar with the NC versus PC
debate, allow me to provide some background. The NC is blah,
blah, blah, Java, blah, blah, trying to screw Microsoft, blah, blah,
no hard disk, blah, blah, Larry Ellison.
The “blahs” are Mr. Adams's, not mine. Larry Ellison is the CEO of
Oracle and a big proponent of NCC in general, and of NCs in particular.
Since the PC industry is driving many of the innovations in both
academia and industry these days, the NC versus PC debate in the
context of NCC will very likely affect you whether you follow the debate
or not. There is currently a tremendous amount of discussion about NCs
versus other types of thin clients. Time will tell which type will be the
most popular, but history has shown that it doesn't usually pay to bet
against the Redmond juggernaut.
SEI Interactive, 6/98
About the author
Scott Tilley is a visiting scientist at the SEI. He works with the Product
Line Practice Initiative in the Reengineering Center, focusing on
transitioning best practices in legacy-system reengineering in a
disciplined manner. Before taking on this role, he was on leave from IBM
and a member of the Rigi project in the Department of Computer Science
at the University of Victoria. Tilley is the author of a 1993 book on home
computing and has over 50 publications. He has a Ph.D. from the
University of Victoria. He can be reached at email@example.com.
The views expressed in this article are the author’s only and do not represent directly or
imply any official position or view of the Software Engineering Institute or Carnegie
Mellon University. This article is intended to stimulate further discussion about this
For more information
Please tell us what you
think with this short
(< 5 minute) survey.