BBS水木清华站∶精华区
发信人: linuxrat (叫我老鼠错不了), 信区: Linux
标 题: Linux定位的新目标?[FWD]
发信站: BBS 水木清华站 (Mon Jan 17 14:58:10 2000)
不知道我的标题是否理解有错. 哪位给纠正一下哦. :)
URL: http://gartner3.gartnerweb.com/public/static/hotc/hc00085653.html
=====Begin=======
Research Note
Strategic Planning Assumption
11 January 2000
A New Approach to Linux Positioning: The Long View
G. Weiss
Hype states that Linux will overthrow NT. A more interesting scenario
shows Linux reinvigorating the Unix market.
_________________________________________________________________
Core Topic
Unix and Midrange Server Vendors and Markets ~ Hardware & Operating
Systems
Key Issue
How will Unix and midrange server vendors and markets evolve?
Strategic Planning Assumptions
By 2002, all four major Unix vendors (Sun Microsystems,
Hewlett-Packard, IBM and Compaq Computer) will position Linux as the
midrange Unix complement to their high-end server strategy but will
not relinquish their positioning of Unix as the strategic enterprise
mission-critical OS environment (0.7 probability).
Through 2003, 90 percent of Linux server deployments will be on Intel
IA-32 (rather than IA-64 or RISC) (0.8 probability).
_________________________________________________________________
The Linux community wants to replace well-established OSs. It can be
done, but enterprise planners take years to redesign, architect,
redeploy and reconfigure systems and change vendor relationships. We
fault the Linux community's claim that Linux will topple Microsoft.
Instead, we present a scenario of accelerating Linux acceptance in
mainstream corporate applications where resistance lingers (outside
the "hot spots" in education, technical computing, consulting and
specific-function network services).
The Unix market spans a variety of user needs. We foresee a Unix/Linux
OS continuum (see Note 1) including a vendor-specific Unix-branded
segment (as today with Solaris, HP-UX, AIX, Tru64, etc.) focused on
enterprise functionality, deep vendor support, extreme levels of
scalability and mission criticality. At the other end of the market is
a low-end "vendor-independent, open-source-driven GNU Linux" in which
the OS and allied applications conform to GNU principles. This market
focuses on appliance-style, shrink-wrapped "Unixlike" servers, many
pre-configured and deployed in infrastructure services. This enables
Linux to move upmarket as demand and technological maturity warrant.
Parts of these markets are evolving and are wide open competitively.
Linux, if allied to Unix, would reinvigorate Unix, especially in
competitive bidding against Microsoft, in low-end and midrange markets
sensitive to price and price/performance. Here, vendor-branded Unix
and vendor-independent "Unix" could evolve to be nearly seamless, with
common APIs, utilities and applications — particularly on Intel
architectures, which account for more than 90 percent of Linux
deployments. The argument for strategic alignment of Unix and Linux
follows.
_________________________________________________________________
Note 1
Unix and Linux Compared
The Unix and Linux Business Models: The Unix-branded vendor business
model is based on high unique value (architectural and system
enhancements for high availability, capacity on demand, quality of
service, performance and scalability, disaster recovery, etc.). The
market is characterized by low unit volume growth at high ASPs. The
Linux business model is based on low vendor-unique value but pervasive
use in the Internet fabric. High replication value comes from
eliminating the high license fee "tax" associated with vastly
replicated sameness.
Common Unix/Linux Building Blocks: Posix, X Windows, C/C++, NFS, BIND,
Sendmail, LDAP, FTP, Telnet, TCP/IP, Java, Apache, Netscape, Unix
shells.
Areas of Unix/Linux Divergence: Device drivers; server diagnostics and
problem resolution; specific architectural enhancements for ccNUMA,
clustering, high availability and failover; special OS functions such
as capacity on demand, resource and workload management, and quality
of service; OS partitioning and remote systems management.
_________________________________________________________________
1. Linux stalwarts have made it a credo that Linux is GNU (see Note
2), to differentiate Linux from fragmented and steeply licensed Unix
originated with AT&T's ownership. However, GNU is too broad and
far-reaching for broad vendor and ISV acceptance (e.g., because of
issues involving patents and intellectual property rights).
Compatibility with Unix is one of Linux's main attractions as a
complementary strategy.
_________________________________________________________________
Note 2
GNU
GNU's Not Unix, or GNU, was invented by Richard Stallman. The reason
for GNU is Stallman's goal of developing an entire OS environment in
which the following principles apply: 1) freedom to run programs for
any purpose, 2) freedom to modify the program with access to the
source code, 3) freedom to redistribute copies of programs (gratis or
for a fee) derived from the open source and 4) freedom to distribute
modified versions of programs.
_________________________________________________________________
2. Linux fits where Unix hoped, but failed, to win: as a commodity OS
for inexpensive Intel servers and desktops. It solves the enterprise
need for budget-tight deployments (e.g., replicated sites, Internet
nodes and embedded devices that can be easily scrapped, swapped and
upgraded).
3. Linux has not yet emerged as a widespread application or high-end
DBMS OS. Linux must show better scalability, performance and worldwide
vendor platform support to attract ISVs. This requires ongoing vendor
hardware enhancements and optimization. As a complementary Unix
strategy, Linux has the opportunity to mature and yet not compete
directly against incumbent Unix, thus satisfying the Unix vendors (see
Note 3). As an example, three-tier ERP applications on Linux IA-32
Intel servers could be an alternative to NT, with the DBMS on Unix
servers.
_________________________________________________________________
Note 3
The Vendor Impact
1. SCO and Solaris (Intel) target the heavily replicated Intel server
market with branded Unix OSs. This model falls apart under the more
attractive Linux positioning presented in this Research Note.
2. Neither IBM, HP, Compaq nor SGI can accept the complementary
strategy, since RISC Unix at the low end will eventually be starved
out by the better price/performance of Intel, although support
requirements will increase. While Unix will survive on Intel, it will
not survive well on low-end IA-32 hardware against Windows 2000, Linux
or Novell.
3. As Linux is increasingly accepted for Unix at the low end and
midrange, RISC-based architectures will come under increased pressure
by 2003, since Intel will fill out both the low and the high end with
IA-64 technology, and the Linux movement will spread OSS applications
across a fuller spectrum.
4. Unix vendors can choose either to provide an Internet version of
their branded Unix at virtually no cost, including OSS-ported
applications, or to turn to Linux as an attempt to weave Linux into
their hardware and software architectures (e.g., a common management
framework).
_________________________________________________________________
4. Linux will not displace Unix wholesale from the high-end market in
the next three to five years because: Unix is well-entrenched legacy
technology, Linux is not easily portable to RISC architectures and
would represent the same kind of fragmentation that Unix has, Linux
would complicate existing OS heterogeneity problems, and Linux
currently lacks a fast-track process to achieve mission-critical
enhancements equivalent to those offered by Unix vendors.
5. The Linux service/support model is limited to the OS and
surrounding infrastructure. This support is usually available in most
enterprises (through Unix and NT experience); most added-value support
is for the OS and future enhancements and upgrades, which vendors can
outsource to Linux distributors and vendors. Linux reliability and
remote support (through its modular nature) will lessen the need for
24x7 service and support, but the burgeoning acceptance of Linux can
still support several quality Linux vendors in addition to the Unix
vendor community (especially where global deployments are required).
Linux and Unix communities would become symbiotic.
_________________________________________________________________
Acronym Key
24x7 24 hours a day, seven days a week
API Application programming interface
ASP Average selling price
ccNUMA Cache-coherent nonuniform memory access
DBMS Database management system
ERP Enterprise resource planning
GNU GNU's Not Unix
ISV Independent software vendor
OS Operating system
OSS Open-source software
RISC Reduced instruction set computer
SCO The Santa Cruz Operation
_________________________________________________________________
Bottom Line: Type C (conservative) and some Type B (mainstream) users
will be less skeptical about Linux if vendors restructure their server
OS strategies to a continuum of branded Unix and vendor-independent
Linux, positioned as a Unixlike enterprise-supported OS. Type A users
(aggressive technology adopters) can retain more independence without
a vendor-led Linux strategy, but at the cost of more time and
resources devoted to the evaluation and selection process. Enterprises
should accept Linux as a low-to-midrange Unix substitute with a
feature/function set similar enough to Unix's to warrant its
widespread deployment, but only once vendors have integrated Linux
seamlessly into their hardware and OS strategies.
______下面这些版权所有声明什么的, 愿意就看, 但我必须保留. :)______
This document has been published by:
Service Date Document #
Unix & Midrange Strategies 11 January 2000 SPA-09-9247
PRISM for Distributed Computing 11 January 2000 SPA-09-9247
Information Technology Management - Distributed 14 January 2000
SPA-09-9247
Information Technology Management Europe - Distributed 14 January 2000
SPA-09-9247
Information Technology Management Asia/Pacific - Distributed 14
January 2000 SPA-09-9247
_________________________________________________________________
Entire contents (C) 2000 by Gartner Group, Inc. All rights reserved.
Reproduction of this publication in any form without prior written
permission is forbidden. The information contained herein has been
obtained from sources believed to be reliable. Gartner Group disclaims
all warranties as to the accuracy, completeness or adequacy of such
information. Gartner Group shall have no liability for errors,
omissions or inadequacies in the information contained herein or for
interpretations thereof. The reader assumes sole responsibility for
the selection of these materials to achieve its intended results. The
opinions expressed herein are subject to change without notice. Please
read the guidelines and policies for GartnerGroup copyrighted
materials. Privacy statement.
=====End=========
--
|======================+========================+====================|
| 以无法为有法 ， | 拳本无法，有法也空； | 我爱GNU/Linux, |
| 以无限为有限 | 一法不立，无法不容。| 因为我爱自由！ |
| | | |
| 截拳道宗师-李小龙 | 意拳宗师-王芗斋 | 土人 Linuxrat |
|======================+========================+====================|
※ 来源:·BBS 水木清华站 smth.org·[FROM: 202.112.168.253]
BBS水木清华站∶精华区