|
IBM Announces Pre-Packaged Blade Superclusters
By Charles King
IBM has introduced the eServer
Cluster 1350, a new pre-packaged solution that combines the company’s eServer BladeCenter technology
with a Linux version of IBM’s Cluster Management Software (CSM) to address
the needs of customers with high-density cluster requirements. The eServer Cluster 1350 can be built with any combination of
IBM’s eServer x335 and x345 systems, and can
combine eServer x345 or x360 storage nodes and an
x345 management node, all running Intel Xeon processors executing at speeds
up to 3.06 GHz. Like other IBM BladeCenter
products, the eServer Cluster 1350 supports up to
84 two-way blades in a single rack. The new product supports IBM’s TotalStorage FAStT200 and FAStT700 storage systems, and
offers an optional fibre switch to simplify SAN
deployment and management. General availability of the eServer
Cluster 1350 with the IBM BladeCenter is June 6,
2003. Pricing details were not included in the announcement.
In the often esoteric worlds of supercomputing and
high performance computing (HPC), the biggest news of the past couple of
years has been the continuing evolution of highly clustered RISC and
Intel-based systems, which are largely supplanting more traditional solutions
in university and research laboratory settings. At the same time, the price/performance
of clustered systems has declined to the point where they have become
increasingly cost-effective for commercial applications in areas including
pharmaceutical/drug discovery, oil and gas exploration, and aerospace and
automotive design. However, even in these areas, supercomputing clusters tend
to be highly customized by vendors who develop them or IT staff members who
deploy, manage, and maintain them, significantly impacting the overall TCO of
such solutions.
How does IBM’s new eServer
Cluster 1350 differ? In a couple of ways. First, by leveraging its BladeCenter technology, IBM has created an exceptionally
dense and flexible solution for clustered supercomputing applications.
Additionally, basing the new systems on Intel’s Xeon processor family, which
boasts considerable forward compatibility, should help ensure that systems
can be easily upgraded, extending their shelf life and ROI. Finally, IBM’s
decision to deliver the product as a pre-configured cluster could
substantially reduce deployment and maintenance costs. Does this mean that
the eServer Cluster 1350 will be a guaranteed hit?
Since the market for what are essentially commoditized supercomputing
solutions remains somewhat rarified, how well these new systems will be
received among HPC novices remains uncertain, but we
expect they will strike a melodious chord among veteran users who appreciate
highly dense and easily deployed solutions. Overall, we believe that while
the market for solutions such as the eServer
Cluster 1350 may be young, it is also promising. If the market rewards IBM’s
proactive optimism, we would not be surprised to see similar solutions from
vendors including HP, Sun, and Dell.
|
|
A Rational Approach
By Jim Balderston
IBM has announced the latest version of its Rational
software tools package, part of the IBM portfolio since the company acquired
Rational in February. The new products are expected to be available in June.
The products, IBM Rational Rapid Developer and IBM Rational Rapid Developer
and Rational Team Unifying Platform, offer a visual development environment
for J2EE development of enterprise applications. The products are designed to
give developers with lesser amounts of J2EE experience the ability to program
enterprise applications in a visual development environment. The tools
support Java, J2EE, Linux, Microsoft .NET, C, and C++ development languages
and offer a number of templates for design and code production, with more promised
for the future. The IBM Rational Rapid Developer is priced at $5,995 or $4,995
to existing customers; the Rapid Developer and Team Unifying product is
priced at $6,995.
This is Rational’s first big news release since
being acquired by IBM earlier this year, and underscores IBM’s decision for
the move. As we move forward into ever more complex distributed heterogeneous
computing environments, the ability to find pre-trained, highly proficient
programmers in the arcane field of enterprise application development remains
a challenge. To further expand the pool of possible candidates, offering a
relatively easy-to-use development tool makes perfect sense. Not only are
more developers brought to the table, but the kind of increasingly complex
applications being touted as the future of computing come within the grasp of
a larger and larger number of companies as a result.
We were heartened to hear IBM officials from
Rational talk positively about the now increasingly popular concept of autonomic
computing, specifically noting that there is real technology in the autonomic
computing stable, not just marketing material. While the specifics autonomic
computing remains an ever-evolving stew that will take at least a few more
years on the stove, Rational’s ability to aide and abet the adoption of
autonomic computing by providing componentry and architectures that allow
autonomic computing-friendly applications to be built by non-alpha-geeks can
only serve to speed the realization of IBM’s vision. Furthermore, these new
development tools that can build autonomic enabled applications establishes a second, crucial phase for autonomic
computing. Up to now, all of the autonomic computing implementations have
largely been a product of IBM’s software and support organizations. With
these new tools we expect to see more third-party applications and services
adopt autonomic capabilities as IBM continues to expand the genre. More
development, more rapid adoption, and a broader selection of solutions for
the autonomic computing environment. Sounds Rational to us.
|
|
HP Financials: The Verdict is Still Out
By Jim Balderston
HP announced its second quarter earnings results
this week, with $18 billion in revenue and $1.14 billion in non-GAAP profits,
$659 million otherwise. Both numbers were up slightly from the prior quarter.
The company noted that it had cut costs by $3.5 billion on an annualized
basis since its merger with Compaq last year. According to HP, enterprise
systems sales grew 3% over the quarter, with near break-even profits, and HP Services grew 2% with profits. Personal
computing sales remained flat for the quarter, but profitable. Printing and
imaging grew 13% over a year ago, but down 1% sequentially. HP said it was on
track for Wall Street consensus estimate of $36 billion for the second half
of the year, and announced it would cut an additional 3,500 jobs on top of
its ongoing efforts to reduce its workforce by some 18,000 employees. The
company noted that it has some $14 billion in cash or cash equivalents on
hand.
HP touted its latest earnings results as yet more
evidence that its hotly contested merger with Compaq was, in fact, justified.
While these numbers show that the company is at least stable for the time
being, we can not help but wonder if the drive to justify the merger – in the
form of bottom line results – is
proving to be a two-front war, with one being largely ignored for the time
being. We also note that as times have become a little tougher, HP has begun
breaking out its earnings numbers by division, something that the company has
not done recently. We suspect the idea here is to highlight good news where
it can be found in the hope of affixing a shinier patina to essentially
neutral results.
And for the time being, the best we can call these
results is exactly that – neutral. HP has made its numbers largely on the
basis of cutting costs – or people – out of the equation. While a torpid IT
spending environment is not helping, we have to wonder if the company is setting
itself up to lose a war best fought with IT innovation while focusing on its
latest battle with Wall Street doomsayers. In short, we do not believe that
HP can cut its way to growth. Instead, it has to begin growing new revenue
opportunities by keeping or exceeding the pace of ongoing technology trends.
With so much energy being placed on winning the opinion battle over the
wisdom of the Compaq merger, we wonder where HP thinks critical new revenues
will come from. It is a good sign, we suppose, that HP services continues to
do well, but the company is going to have to continue to provide technology
that is compelling in the face of competitors – like IBM – that are moving
much more quickly on initiatives like On Demand and Autonomic Computing.
While HP is making noises in this arena they are less distinct than their
rivals’. It also appears that for the near term, HP’s utility-style offerings
will be only viable within HP environments, which is hardly a way to grow
market share and revenues over the long haul. HP needs to remember that
computing environments evolve, and those not keeping pace face obsolescence.
Can HP make the leap forward, catch, and even take a lead in development of
the next generation of computing environments while at the same time reducing
staff? In our mind, that remains the real question at hand, and one that
these financial results at least, have failed to answer conclusively.
|
|
Spin Control: SCO UNIX
Announcements/Fallout Continue
By Charles King
SCO has announced that Microsoft has licensed the
company’s UNIX technology including patent and source code. According to SCO,
the deal will ensure intellectual property compliance across all Microsoft
solutions and better enable Microsoft to ensure compatibility with UNIX and
UNIX services. No financial details of the deal were made public. In
unrelated events, according to media reports SCO canceled its marketing,
business development, and financial resource support for UnitedLinux. SCO
also reportedly resigned its membership from LIVE Linux-Verband eV, a German
Linux association, after the group asked SCO to provide additional
information about its claims that UNIX source code was illegally being used
in Linux. These moves came less than a week after SCO wrote letters to 1,500
global enterprises warning them that they could be liable for using Linux
containing SCO source code.
We live in curious times, when mendacious
bureaucrats and untrustworthy corporate executives alike say whatever they
want, then lie, deny, or change the subject when they are called to account.
Conventional wisdom suggests that such dissembling succeeds due to the slow-wittedness
and short attention spans of the public, but we believe another root cause
exists. The sheer volume of information churned through the media on a given
day often resembles the output from a broken sewer main, so is it any wonder
that most people turn their attention to more pleasant pursuits? However, as
analysts it is our not especially solemn duty to don gumboots, grab a shovel,
and clear away a bit of effluvia to see what lies beneath. Regarding
Microsoft’s licensing of UNIX, we must tip our hats to the folks in Redmond
for playing a brilliant game. At heart, Microsoft believes that business is a
contact sport and that anything or anyone that fails to support the company
adequately (according to the Microsoft lexicon, anyway) deserves to floss
regularly with Bermuda grass. At this juncture, only two serious impediments
stand between Microsoft’s domination of business computing: IBM, which
regards Microsoft as merely one weapon in its larger enterprise arsenal; and
the growing popularity of Linux in corporate IT. Essentially, licensing UNIX
from SCO has allowed Microsoft to prop up SCO’s Quixote-like tilting at IBM,
sow fear, uncertainty, and doubt amongst UNIX and Linux users, and at the
same time wrap itself in the twin flags of customer concern and respect for
intellectual property.
So what is SCO really up to amidst this hubbub?
First, it must be remembered that the company has offered no proof to back up
their source code misappropriation complaints against IBM or cautionary memos
to Linux customers, claiming that doing so would undermine their case. As
such, the company’s actions have left IBM and other Linux vendors and users
in the Kafkaesque position of defending themselves against the most ephemeral
of accusations. It is easy enough to see why SCO is pursuing this course. So
long as it maintains whatever “proof” it holds under lock and key, the
company owns the spin cycle. As soon as the evidence becomes a matter of
public record and debate, SCO will become a bit player in the drama it
brought to the stage. This is may be smart strategically, but also perfectly
in keeping with the company’s recent dismal history. So far as product
development and innovation goes, SCO has, for some time now, been the IT
equivalent of a yokel who could not find a cold beer at a wet t-shirt
contest. SCO’s apparent unconcern for the bridges it is burning among UNIX
and Linux partners suggests that the company is focused wholly on pursuing a
strategy to maximize its net worth prior to acquisition or dissolution. This
may be a good play for SCO shareholders (including stock option-owning
executives), but it is dangerously myopic as regards the larger UNIX and
Linux communities. At least the folks in Redmond will be enjoying a
well-produced and subsidized show.
|
|
|