The Changing Face of Linux

1
The Changing Face of Linux

• Enterprise Linux Forum
• 3 Dec 2002
• David C Niemi
• Adeptech Systems, Inc.
• ltdcn_at_adeptech.comgt
• http//www.tuxers.net/niemi/elf/

2
Outline

• Speaker Introduction
• Why Develop Free Software?
• The Linux Experience
• Why Linux is Different
• Views from the Past (1987-2002)
• Recent Accomplishments
• Current Challenges for Linux
• The Next Five Years For Linux
• The Next Five Years For Microsoft
• Footnotes and Links

3
Adeptech Systems, Inc.
Speaker Introduction
David C Niemi

• System engineer, UNIX and Internet application
  developer, security engineer
• Contel/GTE 1985-1994
• Sallie Mae 1994-2001
• Adeptech Systems 2001-
• Linux contributor and maintainer, 1994-present
• Founder/Director, Tux.Org Inc.

• Based in McLean, VA
• Founded 1995
• Security Analysis
• Network Architecture
• Messaging, Directories
• Beryllium Project an open source application
  integration framework

4
Why Develop Free Software?

• Ideally because you are paid to develop free
  software
• Contribute to a larger effort/collaborative
  development
• Permanence good public source code never really
  dies
• Commercial software lives and dies by its
  commercial viability and corporate circumstances
• Shareware's life is limited by lack of source
  code
• Free software projects make continuous progress
• Experience, learning, resume-building
• Widespread public benefit of free software use
• Fame, recognition by peers
• Get help in maintaining and improving software

5
The Linux Experience

• Linux is technically only the kernel and a few
  utilities
• Distributions include far more GNU, XFree86,
  others
• Linux is also an attitude and an approach
• Organized over the Internet from the start
• International and inclusive a meritocracy
• Idealistic, optimistic, willing to break with
  tradition
• Start with simple but correct code, good
  interfaces
• Rewrite later for performance and new
  functionality
• Increasing use of performance metrics in design
  2 steps forward, 1 step back isn't good enough

6
Why Linux is Different

• Like UNIX in philosophy and interfaces, but
• Implementation is all new old assumptions
  revisited
• Smaller and simpler (at first)
• More POSIX-oriented
• Rapid development, evolution
• More focus on end users MS compatibility
  options
• No single-vendor control -- a level playing
  field for all

• Different priorities from MS technical, not
  marketing
• Modular kernel, services, GUI all separate no
  DLL Hell much easier diagnosis
• Full, modifiable source code
• UNIX vs. DOS semantics case sensitive, forward /
  mount points, virtual files
• Fully multi-user and remotely manageable

7
1987 Precursors to Linux

• ATT has taken UNIX proprietary many licensees
• UNIX overtaking VMS, Sun overtaking Apollo POSIX
  1
• Servers interact with dumb terminals over serial
  lines
• The Internet quiet but rapidly growing research
  network
• GNU Emacs available, work on GCC and other tools
  underway
• Linus Torvalds in high school

• Microsoft MS Windows 2 (not a success),
  PC-DOS 3.3 no antitrust trouble yet!
• OS/2 1.1 with Presentation Manager
• Apple strong 20 market share, 50 gross profit
  sues HP, MS in 1988 over look'n'feel
• Lotus, Ashton-Tate, WordPerfect dominate PC
  applications

8
1992 Linux in its Infancy

• UNIX Wars are finally over RISC CPUs taking
  over high end
• The Internet reaching critical mass
• 386bsd and NetBSD available for free stable but
  hard to install
• Linus Torvalds at U of Helsinki
• Linux distributions appear MCC, TAMU, Yggdrasil,
  SLS in 1992-3
• dozens of Linux users grow to thousands 0.12
  and 0.95 kernels

• MS Windows 3.1 a big hit
• OS/2 2.0 released
• MS Office winning office suite competition
• Apple losing look'n'feel lawsuit (final appeal
  1994)
• FTC Deadlocked on MS antitrust investigation

9
1997 Linux Growing Up

• Microsoft discovers the Internet, maims Netscape
  Linux still below radar screen
• NT 4.0 a success MS Windows 98 nearing release
• 2nd DOJ antitrust probe
• OS/2, NetWare in decline MS rules applications
  market
• Apple near-death experience

• UNIX, Linux, BSD servers boosted by Internet boom
• Linux, Windows NT replacing UNIX workstations
• Linux 2.0 kernel supports SMP and non-Intel CPUs
• Linus Torvalds in California
• Linux distributions Red Hat, SuSE supplant
  Slackware
• Several million Linux users

10
2002 Linux threatens Microsoft

• Internet boom long gone, tech markets still weak
• HP/Compaq merger Intel, AMD CPUs outperforming
  RISC
• Linux distributions consolidating Red Hat and
  Mandrake popular, United Linux debuts
• Linux gains strength on desktop
• Linux 2.4 kernel stable, 2.6 near
• Estimated 18 million Linux users

• MS considers Linux 1 threat carries out
  lobbying, marketing campaigns against Linux and
  other free software
• MS Windows 2000 and XP
• 2nd DOJ lawsuit ends in toothless consent decree

11
Recent Accomplishments
Nontechnical
Technical

• Easy installation processes (distributors)
• Wide hardware support (community and vendors)
• Greatly improved scalability (kernel developers,
  vendors)
• Slick End-user Desktops (KDE, GNOME, Ximian)
• Slicker End User Applications (Sun, Ximian,
  community)
• Antialiasing, TrueType Fonts (XFree86)

• Preinstalled Servers (Dell, HP, IBM, etc.)
• Third-party server and network applications
  (vendors, community)
• Advertising (IBM et. al.)
• Policymaker awareness (community, distributors)
• IT architect awareness (community, vendors)
• Enterprise respectability (result of all of the
  above)

12
Current Challenges for Linux
Challenge
Status Today

• Easy for end users Mac OS X, NeXT prove it is
  possible
• Slick, comprehensive system administration tools
• Common Criteria certification
• Ports of key third-party end-user business
  applications
• Compelling new end-user business applications
• Clustering for business-oriented applications

• Major recent usability gains good basic desktop
  features
• Several disjoint but promising approaches
• Good technical security
• Most end-user applications are MS Windows only
• Groundbreaking scientific and network-related
  tools
• Linux HA project, Beowulf clusters (scientific)

13
The Next Five Years for Linux

• The end user environment is strategic
• Large companies wield great influence IBM, HP,
  AOL, Sun, Oracle, Dell (increasingly helpful to
  Linux since 1998)
• Software-related legal issues entering political
  arena governments making conscious decisions on
  free software Linux could be outlawed, mandated
  or some of each
• Latin America and Europe will turn more strongly
  to free software (including desktop) unless
  discouraged by US
• Free software is a natural political and economic
  fit for China, India, Russia may subsidize free
  software

14
The Next Five Years for Microsoft

• EU antitrust case and numerous private cases
  continuing. Weak consent decree will increase
  scrutiny, not end it.
• US Government central to lobbying efforts
  increasingly out of step with other software
  vendors, foreign governments
• Needs compelling new applications out of its
  research labs
• Forays into new markets e.g. MSN, WinCE, Xbox
  are big money-losers, may become unsustainable
• Strong software DRM could be a costly strategic
  error
• Risky all-or-nothing strategy in order to break
  the MS Windows monopoly, MS Office must first be
  defeated

15
Footnotes and Links

• This paper and related notes and
  updates http//www.tuxers.net/niemi/elf/
• Tux.Org, Inc. http//www.tux.org/
• Adeptech Systems, Inc. http//www.adeptech.com/
• Beryllium application integration framework
  http//www.adeptech.com/beryllium/
• Microsoft Antitrust Case Discussion
  http//www.tuxlaw.org/software-competition/
• The Linux Counter Estimating the number of Linux
  users http//counter.li.org/estimates.php