Librenix
Headlines | Linux | Apps | Coding | BSD | Admin | News
Information for Linux System Administration 

The Network Computer: An opportunity for Linux

Up
vote
Down

Network Computer Terminology
Sometimes you will hear of a network computer referred to as a 'thin client'. While there is some overlap in usage, the two terms originated in opposite realms. 'Thin client' historically refers to a small, simple software client that runs only the user interface for a multiuser application. In an increasingly broad range of applications a web browser is used, without modification, as this client tier.

The term 'network computer', on the other hand, is usually used to refer to a small footprint PC replacement that appears, at first glance, to serve a similar function as the mainframe terminal of years past, but with mouse input and a graphical color display. The fusion of the two terms is probably due to the fact that the network computer includes a thin client in its firmware. While this article is nominally about network computers, some of the points that I will discuss here are actually about the thin client software that is an inherent part of the network computer (NC) itself.

Why Network Computers Will Succeed
The main case I intend to make here is that the network computer, in some form or another, will inevitably succeed on a large scale. This success will be almost totally a business phenomenon, with businesses, governments, and other organizations gradually replacing certain PC functions with NC's over the next decade. The reasons for this upcoming shift are entirely economic. The total cost of ownership (TCO) of network computers will be dramatically lower than the TCO of PC's for most functions. And, for almost all purposes, NC's will be able to reduce TCO while increasing productivity.

The reasons why business will find joining this gradual migration to network computers irresistible are:
  • Security: With no hard drive and no facility to insert malware into the NC, security costs and risks will plummet. Further, in areas with secret or confidential information, the mass extraction of data is much harder without some local writable device like a CD writer or a floppy drive. For this reason, some NC's will either have no USB ports or they will have a facility to disable the use of USB ports from the server.

  • Maintenance: Upgrades to the thin client firmware will rarely be necessary. When it is required, it can be performed en masse from a server. The hardware simplicity and the interface-only function will reduce both the amount of hardware that breaks and the necessity for hardware upgrades.

  • Purchase cost: With a predefined and unchanging function, the hardware components of the NC are few and inexpensive. The processing power need only be enough to rapidly update the display, transfer display-related data over the network, and handle user input. And, of course, no hard drive or CD-ROM drive needs to be present at all.

  • Training: With a simple NC interface and with only those applications available that are necessary for each user's role in the business, training needs will be significantly reduced. And, with a simple and well-tested mini-operating system in read-only memory, and no local configuration data to be modified, user confusion will be a much smaller drag on productivity.

  • Portability: Since any user can log on to any NC and immediately get to only those applications that they use, many problems are solved. User downtime caused by broken equipment, for example, can be almost eliminated simply by keeping a few spare NC's to swap out with broken ones. With no operating system installation and no custom configuration, that operation can be performed in minutes.

  • Backup: Since all data is stored on a server, backups can be performed quickly and cheaply by automated processes on a nightly basis, eliminating the requirement for every computer user to follow data center disciplines in order to properly secure the organization's data and information.

  • Control: With NC's, organizations will finally be able to control what software employees are able to run on their desktops. This has implications for productivity enhancement as well as reducing the company's legal liability.

  • Assured license compliance: With the current PC scenario, virtually every sizable organization has some instances of unlicensed software running on its PC's. This problem can be made to go away completely with NC's since all software can be loaded from centralized servers under the control of a software license daemon, thereby guaranteeing that the license limits are never exceeded.
As you can see, using NC's can reduce -- and in some cases, eliminate -- many of the costs, security risks, and legal liabilities of the ubiquitous PC-on-every-desk paradigm.

Even with these advantages, the network computer will not completely displace the PC. However, it will affect the PC market -- and NC's will certainly replace some PC's.

Microsoft and Network Computers
This rising appeal of NC's is introducing a market that Microsoft and their fellow-traveling entrenched partners do not control. It is for that reason that we will likely see a series of well-funded studies which analyze one aspect or another of the network computer and find it wanting. However, even as Microsoft works to undermine -- or at least delay -- network computers, it will continue its effort to take over that niche as well. Microsoft knows that in the long run NC's will not be stopped.

The Linux Advantage
Microsoft also knows that Linux is uniquely suited to fill the niches of both server and client for network computers. The availability of large applications development projects, a top quality browser, and networked graphical interface technology position Linux -- and the cluster of major projects that integrate so well with it -- ideally to take advantage of this shift to the network computer.

Linux is already proven as a solid and extremely flexible server operating system. Add in the progress already made in embedded Linux and the result almost completes the package.

And, of course, Linux already has the all-important mindshare in a critical mass of organizations to provide a potential customer base for such a project.

All that is needed is for a visionary company with corporate credibility, development capital, and hardware and software resources to take the gamble on implementing a first class network computer and server package with basic office suite functionality and zero licensing costs. That, alone, should be sufficient to create an initial installed base. Once the trend is established, proprietary software, albeit not free, will follow to fill the larger gaps left by the existing free packages. And, of course, that will only serve to increase the rate of adoption.
mail this link | score:8807 | -Ray, October 6, 2003 (Updated: October 18, 2003)
More Sysadmin articles...

Decorate your home or office with Fine Art

Selected articles

Testing the Digital Ocean $5 Cloud Servers with an MMORPG

VPS: Xen vs. OpenVZ

Apple DIY Repair

Librenix T-Shirts and Coffee Mugs!

Linux dominates Windows

How to install Ubuntu Linux on the decTOP SFF computer

Microsoft to push unlicensed users to Linux

Apple to Intel move no threat to Linux

Space Tyrant: A multiplayer network game for Linux

Space Tyrant: A threaded C game project: First Code

Space Tyrant: A threaded game server project in C

Missing the point of the Mac Mini

MiniLesson: An introduction to Linux in ten commands

Tutorial: Introduction to Linux files

Mono-culture and the .NETwork effect

Hacker Haiku

Closed Source Linux Distribution Launched

No, RMS, Linux is not GNU/Linux

Why software sucks

Why Programmers are not Software Engineers

The Real Microsoft Monopoly

Programming Language Tradeoffs: 3GL vs 4GL

Graffiti Server Download Page

Beneficial Computer Viruses

The life cycle of a programmer

 

Firefox sidebar

Site map

Site info

News feed

Features

Login
(to post)

Search

 
Articles are owned by their authors.   © 2000-2012 Ray Yeargin