|
Market Roundup January 30, 2004 HP and Intel Discuss 64-bit for x86 IBM Merges Systems/Technology Groups |
|
HP and Intel Discuss 64-bit for x86
News reports this week claimed that Hewlett-Packard will
announce plans in February to deliver 1- to 4-way Proliant
servers based on AMD’s Opteron processor. According
to the reports, the announcement is expected in early February, but HP may not
deliver the products until later in 2004. In response to the reports, the
company issued a statement saying, “HP acknowledges customer demand for support
from a trusted vendor for x86 extensions technology in certain vertical
segments, where specific price-performance needs exist. HP is currently
assessing our options in this area. We are not disclosing information about
systems, partners, or availability at this time.” In an unrelated event, Intel
President and COO Paul Otellini responded to a
reporter’s question about the company’s plans to deliver 64-bit x86-based
solutions, "You can be fairly confident that when there is software from
an application and operating system standpoint that we'll be there."
Otellini’s statement reiterated similar comments by the company’s CTO Pat
Gelsinger at the Fall 2003 Intel Developer’s Forum. In
addition, a report on CNET claimed that Intel plans to demonstrate 64-bit
revamps of its Xeon and Pentium processors at the Intel Developer Forum in
mid-February. According to the report, the 64-bit extension technology (once
code-named Yamhill) will be labeled CT, and will be available in Intel products
available early in 2005. Intel representatives refuse to comment on the story.
Parsing IT vendor statements is a bit like shoveling
manure in hopes of finding a pony, so let’s roll up our sleeves and see if
either HP’s or Intel’s comments are saddle-worthy. In short, the two companies’
actions exemplify the difference between staying a strategic course and looking
for tactical gain. Intel’s view that 64-bit/x86 capabilities should be delayed
until OS and applications are available has the sensible ring of experience,
but it also contradicts the company’s approach to its 64-bit Itanium processor.
Technical and production difficulties added years to the Itanium’s original
delivery schedule, and the chip hit the market with minimal software support.
This is not especially surprising, since the vast majority of ISVs tend to play
wait-and-see with new technologies before sinking product R&D resources
into them. So why would Intel delay delivery of 64-bit/x86 solutions? For two reasons. First, because it owns the 32-bit x86
market. Since only minor players AMD and Apple are delivering commercial 64-bit
desktop solutions, Intel risks little by avoiding early market jitters and can
pick and choose its targets after the dust settles a bit. More importantly, the
adoption of 64-bit/x86 solutions could further damage Itanium’s continuing
laggardly market adoption. Though the hybrid 32-/64-bit capabilities of AMD’s Opteron are catching users’ attention in a serious way,
Intel likely believes it stands to lose as much or more than it would gain by
offering similarly configured x86 processors. However, should Opteron’s success continue or accelerate, we would not be
surprised to see Intel quietly reassess its 64-bit/x86 product strategy.
So given Intel and HP’s close Itanium partnership (some might say co-dependence), why would HP consider delivering Opteron-based servers that might cannibalize Itanium sales? Because HP is more sensitive to variable market trends than Intel is. The fact is, a server sale is a server sale to HP, whatever processor is inside the box, though the company’s positioning of x86 extensions for specific vertical markets (which mirrors IBM and Sun’s positioning) is designed to mitigate some risk. Additionally, if HP does not have a 64-bit x86 solution to offer, customers who want such products will go to IBM and Sun to get them. Long term strategies aside, manufacturers are judged on their quarterly sales results. Considering the ongoing woes of its enterprise products division, we would not expect HP to do anything that might push their clients into the waiting and eager arms of the competition. This is why the CNET story on Intel’s reported 64-bit x96 demo plans is so intriguing. If the company does perform such a demonstration, it would take some heat off of both Intel and HP, and spill a bit of wind from AMD’s sails. While AMD and its Opteron partners would have the market to themselves for another year or so, the time would be clouded by the impending arrival of Intel’s new competing solutions. If there is truth to this story, and we believe there is, the question will be how Intel and HP will try to position a 64-but x86 offering so it does minimal damage to Itanium, and the answer will be found in whether or not the market buys that positioning.
IBM has announced it is expanding its technical and
marketing programs for ISVs as an ongoing effort to support what it called the
growing adoption of Lotus Workplace by ISVs. IBM said it has already signed up
more than a dozen ISVs to offer Lotus Workplace as part of its ISV Advantage
program, which has more than 150 ISVs in the fold to date. IBM said it is
targeting Lotus Workplace through its ISVs to the mid-tier market and industry
verticals as a collaboration tool. The ISV efforts include offering a number of
components and testing opportunities to ensure application compatibility with
Lotus Workplace, while offering a qualification program for interested ISVs
called Ready for IBM Lotus Workplace. The program will
be available by the end of Q1 2004.
Given IBM’s going push to use ISVs as the entry point into
the mid-tier market, no one should be particularly surprised that Lotus
Workplace would become a part of this ISV-centric strategy. IBM pulled out of
the application business four years ago and instead has decided to use the ISV
community — which owns both the vertical expertise and the relationship with
customers — to move IBM products and services. We have noted before that such a
move is not only sensible; it proffers the opportunity for IBM to reap substantial
profits from the mid-tier market. In short, IBM is going with the flow of
existing market conditions instead of trying to shape reality more to its
liking.
Lotus Workplace is a collaborative environment that includes email, messaging, and web browsing through a single application footprint, a browser, and is positioned in our mind directly against Microsoft’s Exchange and Outlook products. As IBM and Microsoft square off in the mid-tier market, it is essentially that Big Blue puts a viable alternative to these Microsoft offerings in their ISVs’ toolboxes, and in such a way that makes it compelling not only to the ISV, but to the end customer as well. Given the fact that many ISVs face having to make the choice between platforms from Microsoft or IBM, the fact IBM can sweeten the pot with an alternative to Exchange and Outlook may well be the tipping point for many ISVs going forward. Clearly, IBM needs to continue to evangelize Lotus Workplace, especially in the mid-tier market, which may still associate Lotus products with large enterprises and huge installation headaches. To that end, IBM is pursuing a number of co-marketing and visibility activities for Lotus Workplace. Given the fact that many mid-tier companies are seeking to minimize the number of vendors and ISV partners they interact with, the success of IBM’s messaging efforts for Lotus Workplace could be crucial to the overall success of its mid-tier enterprise push, and to the success of its ISV partners as well.
IBM Merges Systems/Technology Groups
According to news reports, IBM will merge its Systems and
Technology organizations into a single entity. The new Systems and Technology
Group will be co-led by senior vice presidents William Zeitler and John Kelly
III, who previously led the separate organizations. The move comes after the
two groups’ eighteen-month “One Team” effort to align strategy around the POWER
processor platform and the company’s new semiconductor fabrication plants. The
new group will issue a single set of consolidated financial results, and
Systems group business objectives will be based on IBM cost, not internal
pricing from the Technology group..
Group or operational mergers are seldom the stuff of IT
legend, but occasionally announcements like those from IBM offer something more
substantial to chew on. Rather than focusing on mundane issues such as product
branding, the Systems and Technology groups’ merger resonates with long-term
strategic issues. Most importantly, it reiterates the importance of POWER
processor technology to IBM. The fact is that no other IT vendor delivers a
chip architecture that offers the sheer depth and breadth of POWER, which
scales from embedded systems and the desktop, to front-end servers and backend
datacenters. Intel (and by turns, HP and Dell) develops discreet chip
architectures and technologies for specific applications, while Sun focuses its
own efforts exclusively on the 64-bit space. Only AMD’s extensions of x86
architecture (with Opteron and Athlon64) offer some
small resemblance to IBM’s approach. In addition, IBM’s recent delivery of the
POWER-based JS20 blade server and the impending launch of its high-end POWER5
processor suggest that the architecture will be under the microscope for much
of 2004, so coordinating the Systems and Technology groups’ POWER efforts makes
eminent practical sense.
Just as important is creating a structured, nurturing environment to boost the IBM Technology group’s success in product and OEM development. Far from being simply an adjunct to IBM Systems, the Technology organization’s efforts have led to the increasing influence of the POWER well beyond the company’s Big Blue walls. The PowerPC 970 core that drives IBM’s own blade servers is also the heart and soul of Apple’s G5 Macs and the company’s well regarded Xserve G5 servers. In addition, given previous deals with Sony and Nintendo, and with Microsoft’s recent decision to move its Xbox consoles to POWER, enterprise vendor IBM stands to become the defacto leader in providing processors for consumer game consoles. The time, money, and resources IBM has put into its semiconductor fabrication operations, including its new fully automated, “touchless” 300mm manufacturing facility, position the company to continue this leadership. Finally, we expect that IBM stands to gain some balance sheet benefits from the merger, since the Systems group’s purchases from IBM Technology will be based on IBM cost rather than OEM pricing models. This is not likely to translate into huge financial gains, but in a business where margins are thinning, and the distance between profit and loss can change in a heartbeat, every penny matters.
Clues to and from the Clueless
The Department of Homeland Security has announced a new
“cyber-alert” system to warn citizens of threats to computers and the networks
they use. The announcement comes on the heels of the outbreak of the Mydoom
virus, which has clogged networks with more than 100 million emails in its
first thirty-six hours of release into the wild, an event the FBI is now
investigating. The new DHS program will allow citizens to sign up for email
alerts and technical bulletins concerning viruses, computer security issues,
and ways by which people can combat these threats.
Sigh. While we recognize that the Internet has become so
mainstream that it now is fodder for election-year posturing, this announcement
will do little but offer some level of assurance to those citizens that are
largely unaware such simple precautions as antivirus software, and the fact
that most of these applications offer not only automated updating, but the
option of receiving email security updates. In other words, we suspect that
these DHS-spawn security notices would arrive in most people’s email inboxes
well after the first missives arrived carrying the threat in question.
This announcement also ignores the fact that the leading antivirus companies have extensive networks of sensors that allow them to track and quantify the spread of viruses worldwide, networks that have been in place and highly effective for years. These networks allow security vendors to properly access threats, and, alongside non-commercial security clearinghouses, set the alarm at the appropriate level due to lengthy experience and a sizable pile of realtime data. No doubt the DHS will rely on similar expertise to issue its warnings, which will, we suspect, be a bit time-stale as they reach their intended recipients. While we are bemused at this election-year posturing, we do think that such efforts have a long-term counter-productive nature, as the IT industry faces growing threats to the network and its component parts and struggles to find ways with keeping ahead of the bad guys. Federal officials have threatened to mandate new security requirements, creating not only anger but open hostility from the private sector. This effort to remind people that the DHS is doing something about Internet security will be seen in knowledgeable quarters as nothing so much as a press release, which speaks either to the incredible cluelessness of the DHS when it comes to Internet security or to the fact that DHS has no responsible or effective thoughts on the matter, leaving them with little authority to dictate terms and conditions of security requirements to an industry that is clearly way ahead of the curve on this matter. What’s next, Queer Eye for the Straight Guy cast doing a makeover on the nation’s color-coded warning system?
BlueArc Titan and the
Future of Storage as We Know It
BlueArc has launched their latest NAS product, Titan, as
a high-performance and scalable storage system for users who want to simplify
their existing environment, use a NAS architecture, and not sacrifice
performance in the process. BlueArc is a start-up who has had significant
growth in the specialized arenas of media and Internet, healthcare and life
sciences, and other high-performance environments where streaming of large
volumes of data is more important than time-to-first-byte. While the original
products were designed with these specific environments in mind, it turns out
that the architecture competes respectably with offerings from mainstream
competitors such as NetApp and EMC, and the new
product is designed to help them grow beyond their traditional strengths and
grow market share in larger segments.
Traditionally NAS product releases are only marginally
more exciting than a new tape product. At the same time, this series of
announcements may actually be worthy of some deep thought. BlueArc’s Titan is
yet another in a newish breed of products, including NetApp’s gFiler, EMC’s Celerra NS600G or NS600GS, and the brand-new IBM NAS
Gateway 500, that form the intersection or melding of so-called SAN and
so-called NAS products. More importantly, like the Cisco MDS 9000 family of
multi-layer switches, these products are representative of a move to drive
storage products away from discrete boxes connecting a subset of storage across
a single protocol to a consolidated and distributed information environment
straddling the cloudy alphabet soup of DAS, SAN, and NAS. In essence once these
controllers and switches are in place, data can go anywhere and talk to
anything regardless of where it was. Or at least it will once the vendors
manage to turn their respective roadmaps into reality. Which means that while
storage loses the acronym war to the telecom industry, it will win the battle
of making storage cheaper and easier to manage, and will expand and presumably
move us closer to the promised land of high utilization and low administration.
Of course there is another irony lurking. The systems vendors spend an awful lot of marketing energy convincing us that they, IBM, Sun, or HP (fill in the blank) are the leaders in technology innovation. And yet, in the storage space the innovators and market leaders such as NetApp, EMC, StorageTek, and Veritas or clever startups like BlueArc continue to be the specialists. Perhaps it has something to do with the fact that many system vendors’ storage products are actually OEMed from one of the industry specialists and therefore not given the respect (or dollars) they might otherwise garner in the internal war of R&D funding. Or, maybe system vendors don’t really believe that their clients view storage at the same strategic level as application processing. At any rate, good enough seems to be as good as it gets. In the end, we wonder what it is about storage that continues to elude the systems vendors.