|
|
Windows, Linux, TCO, and Sponsored Research:
Jambalaya or Junk?
By Clay Ryder
IDC has released to the media a TCO study that
apparently bolsters the argument that Linux is more expensive to administer
than Windows, which in turn can make Windows a more cost-effective server
solution. The Microsoft-funded study queried 104 North American companies and
evaluated costs for networked server computers supporting 100 users over a
five-year period. The servers were used for various tasks including file
sharing, firewall software, print server, and security server. The report
indicated that the largest cost gap was in security servers, where Linux
systems racked up $91k vs. $70k in costs over the five-year timeframe. Print
server applications were Linux $107k vs. Windows’ $87k; file sharing, Linux $114k
vs. Windows' $99k; and Web serving Linux $31k vs. Windows' $32k. IDC also
indicated that administration costs represent 62% of TCO, which is sizably
higher than initial software purchase costs of 4.6% and initial hardware
costs of 4.4%.
One would have to be Rip Van Winkle to have not
heard the ongoing marketing salvos dedicated to the “Which is cheaper, Linux
or Windows?” question. In addition, the same fit of narcolepsy would be
needed to ignore the mathematical or physical slight of hand that often takes
place in all too commonly bastardized Total Cost of Ownership studies. The
cherries on the top of this mess are ongoing news stories about analytically
challenged individuals whose objectivity may have been influenced by the lure
of investment banking or consulting dollars from the very players they were
charged with analyzing. Since TCO studies are arbitrarily defined, at best,
the definition of what is an important cost factor and its relative ranking
is up to the study’s creator. Since there is no one universally accepted
standard, it is quite easy to stack the odds in favor of any given result
while apparently remaining objective. For example, software usage among
businesses is anything but static as suggested by the IDC study. There are
few people or organizations using Microsoft-based software that have remained
at the same version level for five years — few would be running the software
today that they installed in late 1997. In addition, why would those who have
just bent over and signed up for Microsoft Licensing 6 want to stay on a
single software version for five years? One might also ask which Linux
business applications IDC used for comparison that have run unaltered during
to same period. But all this aside, the greater issue at play here is the
role of sponsored vs. unsponsored market research reports and how they are
represented, used, and abused by the vendors, researchers, and the IT
industry at large.
While the publishing of sponsored market research is
nothing new (there are white papers aplenty to attest to this), one needs to
ask whether they were they written because the research company found the
topic too hot to pass up or because they were paid for by a vendor with a
particular viewpoint to push. When presented as such, said reports represent
a view that (hopefully) the analyst agrees with that which he/she has been
paid to codify and explain for the benefit of the industry, the sponsor,
potential customers, and the analyst community. (Hey, room and board aren’t free.) But given the inherent subjectivity of
TCO, the lure of revenue in economically depressed times, and the general
hunger of market players for any economic boost, the temptation to be placed
in compromised positions is all too easy to come by. While IDC indicated that
Microsoft paid for this study, other analyst firms have played more loosely
with the facts (and funding sources) when guarding the proverbial hen house
of objectivity. With all of us who play in the IT marketplace inextricably
intertwined with one another, it is in our best interests to come clean on
this and other comprising positions such as selling to both vendors and end
users and then claiming no conflict of interest in making product
recommendations. In light of this, The Sageza Group last spring ended the
practice of distributing sponsored research alongside our self-motivated
research by creating an entirely different family of products called Sageza
Snapshots, which encompass all of our sponsored works. By discrete branding,
we seek to maintain a degree of separation between the point of view that we
call our own and that which we agree with, but might otherwise not have
published. Will other analyst firms take this issue to heart? We do not know,
but for the good of IT market overall, we believe it is time that all analyst
firms did.
|
|
|
IBM Introduces Linux-Ready p630 eServer
By Charles King
IBM introduced this week a Linux-ready version of
the company’s p630 eServer, a POWER4-based server available
in 1- to 4-way configurations. While the p630 is the company’s first 64-bit
pSeries server dedicated to Linux, it can also support AIX5L or a combination
of Linux and AIX5L in logical partitions. According to IBM, pricing for a
Linux-enabled p630 is up to 44% lower than similarly configured Linux-enabled
HP servers with Itanium2 processors. Through its Linux Technology Center, IBM
offers open source OS features that leverage pSeries systems management,
clustering, Internet services, and security. In addition, IBM plans to offer
middleware including WebSphere, DB2, and Tivoli management software for the
new server. Express Configuration versions of the p630 are currently
available with one, two, or four processors and up to 8GB of memory in rack and
tower configurations with prices beginning at $15,577.
When the p630 was originally launched in June, we
saw it as incorporating key elements of IBM’s larger strategy of migrating
higher-end (i.e., Regatta) server technologies into its lower-end products to
increase their overall business value. While the obvious target for the
original p630 was Sun’s Sun Fire V480 server, IBM is shifting its sights with
the Linux-enabled p630 to HP’s Itanium2-based rx5670. While this is entirely
practical, it also demonstrates the value proposition IBM offers and
balancing act it must follow in the enterprise server market. On the value
side, IBM’s two main rivals, Sun and HP, are pursuing different yet
essentially similar evangelical paths that extol the virtue of their Intel or
SPARC processors. Though the two companies’ motivations differ, the message
is the same: that nearly all of an enterprise’s computing needs can be
satisfied with a single hardware architecture. IBM, on the other hand, is
pursuing a more generalist approach that seeks to develop, position, and
deliver whatever computing technologies are most appropriate for a given
task. IBM’s balancing act, of course, lies in the fact that at the same time
the company is pressing the case for its POWER-based pSeries systems, it is
also developing and delivering Itanium-based systems through its xSeries group.
So what does all this, including the new
Linux-enabled p630, mean to IBM and its customers? To begin, while the 64-bit
Linux world is in an early state with limited applications and resources
available compared to other 64-bit platforms, corporate interest in open
source solutions continues to grow. From that standpoint, the p630 provides
IBM’s substantial base of AIX customers an affordable, flexible Linux/AIX product
for departmental or other server environments, or a development platform for
open source enterprise solutions. The preconfigured Express versions of the
p630 also offer customers substantial financial incentives to stick with or
even explore POWER architecture solutions instead of Itanium-based products.
That may, in fact, be the key to understanding IBM’s strategy behind the
p630. The company made a more sizeable commitment to Linux than the
competition, and the arrival of increasingly powerful Itanium processors from
Intel and larger numbers of enterprise solutions from a variety of vendors
and ISVs means that the 64-bit open source games are about to begin in
earnest. At heart, the p630 provides further evidence of the seriousness of
IBM’s intent to deliver Linux solutions across multiple products and chip
architectures. It cannot be described as an easy path to take, but it is one
that most of the company’s rivals will find difficult or even impossible to
follow.
|
|
|
AOL: They Had To Say
Something…
By Jim Balderston
AOL Time Warner has announced a new set of strategy
initiatives designed to reverse declines in its 35-million member Internet
unit, America Online. America Online plans to begin offering premium content
to its members, much of it drawn from other AOL Time Warner properties, and
will offer premium services like online games for additional fees. The
company also said it was abandoning plans to make subscribers sign up for a
$54.95/month broadband service and instead is trying to offer AOL for
$14.95/month to people who subscribe to other broadband providers, much the
way HBO is offered as a premium service over cable systems. The company also
stressed that it was seeking to more aggressively cater to customer wishes,
including discontinuing third party pop-up ads, a move the company announced
last month. In announcing the strategy, AOL executives acknowledged that they
did not catch the first wave of broadband adoption, had not been a very
cooperative partner in trying to create third party deals with broadband
providers, and, in fact, have not maintained the proper relationship with its
35 million subscribers. The company also announced that it expected online
advertising revenue to be down by 50% in the year 2003.
AOL has been pummeled over the past year or so as
its much hyped merger with Time Warner has fallen almost completely flat,
with AOL — the Internet darling with the vision to see into the future —
instead looking more and more like a stodgy company that completely missed
the latest phase of the Internet revolution. Throw in an accounting scandal,
an increasingly dissatisfied user base, and the fact that the company expects
its once majestic online advertising revenues to
plummet by one half in this upcoming year and you can only reach one
conclusion about this flurry of new strategies: AOL had to say something.
Anything.
But what AOL outlined is not the stuff of a reborn
Internet titan. Broadband is not only becoming more widespread, the process
of hooking up such a connection is becoming so easy that many people’s
mothers — or grandmothers for that matter — can walk through the connection
wizards alone or at minimum with the help of support staff. Furthermore, when
one adds in the thought that broadband connectivity is becoming much more
widespread, the fact that there is always a friend, neighbor, or relative
ready to assist in the process makes many people comfortable with the
process. In short, America Online’s primary selling point of being really
easy to use is losing its luster. We also suspect that offering premium
content may also fall flat, since there is no shortage of online content for
consumers. In many cases what America Online is eyeing as premium content is
transplanted directly from the Time Warner stable. If one assumes that AOL
Time Warner is willing to cannibalize its content to save America Online — a
big if — one has to ask: what’s the real difference between Time and
Newsweek? Is there really a premium there? Finally, we believe that America
Online‘s intrusive pop-up ads, aggressive selling of customer information,
and increasingly hostile attitude toward subscribers in the past few years is
finally catching up with the company. The word is out that this is no longer
your grandmother’s America Online, and that the old girl may be far better
off with another ISP. In short, we suspect this bold new set of strategy
initiatives was little more than an attempt by America Online to buy a little
time, while it tries to think up something that might actually work.
|
|
|
HDS Goes After Mid-Market Storage
By Charles King
Hitachi Data Systems (HDS) has announced the new
Freedom Storage Thunder 9500V Series storage systems. According to HDS, the
Thunder 9500V products are aimed at customers looking for high performance,
large capacity storage systems in a small footprint that are optimal for
small- to medium-scale consolidation. The new systems come in several
configurations. The Thunder 9570V is a rack mountable system with up to 4GB
cache and four 2GB/sec ports, and supports up to 224 drives (32TB) of storage
per system. The Thunder 9531V, 9532V, and 9533V systems are available in
preconfigured, specially priced storage decks ranging from 360GB to 1TB in
capacity, and can be upgraded to the 9570 system. The Thunder 9500V products
feature a virtualization assist layer that HDS says allows administrators to
mix heterogeneous servers on a single port and provide security to the LUN
level through Host Storage Domains. The Thunder 9500V Series will be
available in January 2002. No pricing information was included. In a separate
announcement, HDS and Network Appliance said the companies have signed an
agreement to market and sell enterprise NAS solutions. Under the terms of the
agreement, HDS will offer NetApp enterprise NAS gateway solutions for Hitachi
Freedom Storage environment managed by HiCommand Management Framework tools.
The new NAS gateway will be available for Hitachi high-end and modular
products starting in the first calendar quarter of 2003.
These HDS announcements hardly qualify as
surprising, though the NetApp agreement offers an interesting view into the
current nature of IT partnerships. HDS’s new
Thunder 9500V series is probably best considered a late arrival in an
increasingly crowded market. HP’s StorageWorks and IBM’s FastTseries are well
ensconced here, along with StorageTek’s D Series machines. But the three have
also been feeling pressure from EMC’s CLARiiON CX Series (CX200, CX400,
CX600), which has been increased by EMC’s partnership with Dell (which sells
the CX200 through its own organization and channels). Given those
circumstances, how does the Thunder 9500V Series measure up? Not too badly,
though HDS’s shining capacity claims (32TB per
system) were quickly eclipsed by EMC’s capacity bump in the CX600 to 35TB per
system earlier this week.
Technical leapfrogging aside, what chance of success
does HDS stand with the Thunder 9500V? Fairly slim, we would say, but not
just because the space is already so densely populated. The fact is that
while HDS has made its reputation with high-end storage products aimed
squarely at the large enterprise market, just how it is perceived among
mid-market customers is less certain. Additionally, its OEM agreements with
both HP and Sun (who sell its high-end 9900Series machines) will not offer
HDS much joy here since both companies offer their own mid-market products.
Which brings us to the curious agreement with
NetApp. From a conventional strategic standpoint, HDS and NetApp joining
forces could potentially opening new markets for both companies. NAS kahuna
NetApp should learn HDS’s secret large enterprise
handshake and HDS might become a household name among NetApps myriad SMB
customers who are ready for the storage big leagues. However, we wonder if
the focus and functionality of the partnership is too narrow to do either
vendor much real good. Over the past couple of years, a number of vendors
have increased their product offerings in order to successfully sell further
into existing customers. In the case of EMC and Dell, two vendors with quite
disparate client bases and solution sets created a structure that has
broadened and profited both. Overall, though the HDS/NetApp deal could
provide the principals some discreet assistance, we do not believe it will
deliver a similarly broad or positive dynamic.
|
|
|
PeopleSoft Asserts Its Web
Services Position into Business Intelligence
By Myles Suer
PeopleSoft has announced the December 2nd
availability of PeopleSoft Enterprise Warehouse 8.8, which provides pre-packaged
analytic content, a portal interface, and a Web services interface.
PeopleSoft indicates 8.8 can reduce the cost and time for deploying,
managing, and delivering enterprise analytics. More importantly, PeopleSoft
claims it provides immediate access to a consistent view of enterprise data.
Its Enterprise Warehouse consists of an operational data store for offline
reporting, a data warehouse for hosting analytical applications, and
pre-defined data marts for multi-dimensional analysis. It also includes more
than 1,200 pre-defined business metrics. Pricing information was not
announced.
We believe that Web Services are a very important
technology for Business Intelligence. As important, analytics have no value
without data. As Web Services become deployed, data warehouses will be able
to integrate more the enterprises’ vast data stores. This is a notoriously
difficult area and broad claims can be damaging to the credibility of the
information management technology as a whole. The practical issues are not
exciting but are key to a useful deployment. These issues include data
quality, data fusion from multiple sources, formulating relevant analytical
searches, and timely and efficient distribution of the results to front-desk
people. It should be noted that PeopleSoft’s analytic tools are not home
grown and were acquired from its relationship with Cognos. Nevertheless, we
are skeptical about its claims regarding immediate access or for it obtaining
a business advantage from its early deployment of Web services. Data
cleansing is a huge part of creating a data warehouse. It does not matter
whether the company is only utilizing PeopleSoft software. Each data silo
tends to have different data definitions and inconsistency. As we see things,
Web Services create a much improved way to connect between the enterprises’
disparate databases and way to discovery at a certain level data definitions
and file formats.
Given this, we are skeptical that Web Services on
its own will give PeopleSoft a leg up on the Business Intelligence
competition. The name of the game is now a full portfolio of information
management applications from data capture and storage to fused and analyzed
information presented through portal-like interfaces to business operative
throughout the enterprise. This is only achievable through partnerships, and
today this market includes a mix of specialists and equally integrated
players to PeopleSoft including but not limited to Business Objects,
Informatica Corporation, E.piphany, Inc., Hyperion Solutions Corporation,
MicroStrategy, Inc., Oracle Corporation, SAS Institute Inc., and SAP AG.
Standing tall with your head held high on the technological moral high ground
is one thing, but standing alone in today’s market is a completely different
matter.
|
|
|
There Is No Excuse for Poor
Management! Peregrine Sells Remedy to BMC
By Myles Suer
Peregrine Systems has announced the sale of its
Remedy business unit to BMC software last week for a greatly discounted $355 million
in cash. The sale is part of Peregrine’s efforts to reorganize under Chapter
11 bankruptcy protection. The price fetched is a fraction of the price paid
in August of 2001 when Peregrine Systems paid $1.08 billion to buy Remedy,
which at the time was a free-standing corporation.
Although we cannot attest to the reasons for
Peregrine Systems’ financial management woes, we can attest to its poor
management and business strategy. When Peregrine went public in 1997, it
defined its business as Enterprise Service Desk Software. This consisted of
internal help desk software and a series of IT management tools. At the time,
the company’s major competitors were Vantive (now part of PeopleSoft) and
Remedy. During the years that followed, Remedy stayed focused on building
software for IT service management and customer support help desk software
(including bug tracking software). Most mergers fail to work because there is
not enough synergy. In this case the potential did exist to claim a
significant position for the internal and external help desk software market.
However, in contrast to Remedy, Peregrine continued to expand the markets it
serviced to include change management, inventory and configuration
management, work management, and asset management, among others. In 2000,
Peregrine changed the definition of its business to reducing the fractional
cost of doing business for its clients’ organizations. With this action, its
products and services addressed three principal domains: infrastructure
resource management, employee relationship management, and e-commerce
technologies and services.
By the time Peregrine acquired Remedy it was going
in a hundred different directions. How Remedy fit into the mix was no longer clear, and worse yet Remedy’s 9,850 customer sites
including over 60% of the Fortune 100 were never given a plan for how the
integration would work or which software would be continued or be phased out.
With no concrete and clear vision, the two companies were never really
integrated while many of Remedy’s top management walked out the door. This
business tragedy demonstrates how important it is for businesses to have a
clear targeted strategy and the will power to stick to it despite what other
seemingly cool things may present themselves. In this case, the merger may
have been more successful if Remedy instead had acquired the disjointed and
overextended Peregrine at a later date.
|
|
|
Bridging the Gap between Business and
Technology: IBM Announces On Demand Innovation Services
By Jacques Halé
IBM recently announced the formation of a new group
within the Research Division called “On Demand Innovation Services” to be led
by Peggy Kennelly. The new group of about 200 researchers, with a budget of
over $1 billion over the next three years, will work in partnership with
IBM’s Business Consulting Services. The group will focus on Advanced
Analytics (business modeling), Business Process Transformation, Information
Integration, and Experimental Economics.
At first glance, this initiative seems to be in conflict
with the recent acquisition of the PwC consulting arm, now called IBM
Business Consulting Services (IBCS). IBM purchased PwC because it did not
have the business process capabilities in IBM Global Services (ISG), which is
essentially a project; IT integration; and resource management organization.
However, on reflection, the “On Demand Innovation Services” group can be seen
as a bridge between the technical and the business sides and further evidence
that IBM is reorganizing its resources in order to exploit the re-emerging
e-business market with services. Most vendors concentrate on the IT aspect
and are neglecting the need to support business executives in adopting the
e-thing. What is needed is a non-IT approach to process modeling and design.
All the new Web Services standards, such as Web Service Architecture and
Specification (from W3C), BPMN, BPML, WSCI, and ebXML
are very technically oriented. The subjects proposed for the new group could
help to bridge the gap between IT and business, looking at the design and
economics of processes, and could restore the image of IT in the eyes of the
executives tired of the “value-added” of IT. We hope that the term
“Experimental Economics” does not frighten them off again.
|
