|
|
IBM and SGI Both Announce New
Supercomputing Offerings
By Charles King
IBM has announced Supercomputing On Demand, a new
option that allows company customers to access POWER- or Intel-based
supercomputer clusters and pay for processing power based on the required
capacity or duration of use. According to IBM, the new service, which is part
of the company’s greater On Demand initiative, will be especially valuable in
industry sectors such as petroleum exploration, digital media, and life
sciences, where supercomputers are necessary at certain times in product
development cycles but are often otherwise underutilized. The first company
to access IBM’s Supercomputing On Demand is PGS Data Processing, a division
of Petroleum Geo-Services, for an advanced seismic imaging project in the
Gulf of Mexico. To supply processing power for the new service, IBM will
build a grid consisting of hundreds of IBM eServer p655 systems and a massive
Linux cluster of Xeon-based eServer x335 and x345 systems. The initial IBM
supercomputing hosting facility will be based in Poughkeepsie, NY, with other
national and international facilities to follow. In an unrelated event, SGI
announced the availability of the new Altrix 3000
family of servers and clusters, systems which combine SGI supercomputing
architecture, the company’s NUMAlink system
interconnect fabric, Intel Itanium 2 processors, and Linux. Each Altrix 3000 node can run a single Linux OS image with up
to sixty-four Itanium2 processors and 512GB of memory, and can be scaled up
to hundreds — and eventually thousands — of processors. The entry level Altrix 3000 server with four processors and up to 32GB of
memory will be available in the first quarter of 2003 beginning at $70,176.
In a sense, these two announcements offer glimpses
into current and emerging supercomputing thinking. While the utilization of
Itanium 2 processors makes the SGI announcement something of a novelty, the Altrix 3000 also offers the impressive performance and
scalability projections SGI has often been noted for. In fact, while SGI has
had more than its share of problems in recent years, the company has retained
its aura of engineering leadership in high-end technical computing. The
question in our minds is whether or not the market will really care. SGI is
targeting the Altrix 3000 family at the usual
supercomputing suspects (research labs, universities, and
government-sponsored agencies) that can afford high end products. But the
performance metrics landscape tends to change rather quickly, and solutions
that eclipse today’s speeds and feeds statistics often look downright anemic
in six months’ time.
To our way of thinking, the real news in
supercomputing and high performance computing is the continuing migration of
these technologies from lofty research labs to day-to-day commercial
applications. Part of this evolution is being driven by simple
price/performance issues, along with the continuing development of high speed
networking and clustering technologies. As a result, solutions that would
once have unaffordable by any measure in commercial enterprises are now seen
as cost-effective in the extreme. However, commercial supercomputer
installations remain of greatest interest to massive enterprises such as automobile
and aircraft manufacturers with scores or even hundreds of products in
development. Where does that leave smaller players who would enjoy
supercomputing capabilities but do not require full time access to these
solutions? That is the sweet spot IBM’s initiative is targeted at. By
leveraging the company’s notable supercomputing expertise and products, IBM
will not just benefit customers who need limited access to these systems, but
will also likely help seed the market for future sales of standalone solutions.
If this initiative is successful, we also believe it will help validate IBM’s
recently announced On Demand strategy and could inspire similar initiatives
among the company’s supercomputing competitors.
|
|
|
Expecting More, Getting More
By Jim Balderston
The Pew Internet and American Life Project has
released its latest report entitled Counting
on the Internet, which measured how much Americans expect, and then
actually find, information they are seeking online. The report’s findings
note that when it comes to government agencies, 65% of all Americans expect
the Web to have that information, where 82% of Internet users expected to
find the government information while 39% of non-users held the same belief.
In the realm of electronic commerce, 63% of Americans expected a business
will have a Web site that gives them information about a product they are
considering purchasing. Nearly 80% of Internet users held this expectation,
as did 38% of non-users. Sixty-nine percent of Americans expect to be able to
find reliable, up-to-date news online; 85% of Internet users agree as do 43%
of non-users. Two-thirds of Americans expect to find reliable health
information online, with 81% of Internet users feeling this way, and 45% of
non-users. When combining these four areas, Pew found that 84% of Americans
expect to find the information they seek in one of these four categories,
with 97% of Internet users in agreement with this expectation as well as 64%
of non-users. The report also notes that by September 2002, more than
two-thirds of all Internet users in the U.S. had been online for more than
three years, and that this is contributing to ever-increasing expectations
toward what will be available online.
While the developments outlined in the latest Pew
study should not come as a big surprise, there is more at work here — and to
take note of — than simply the maturing of the Internet user and the
mainstreaming of the medium in American life. There are lessons for decision
makers and strategists at enterprises that are planning to — or claiming to —
make the Internet a key element of their future strategies for selling
products or providing information.
As we see it, the real lesson here is that Internet
users are going to go where they want to go, and as they gain more experience
they gather the tools necessary to navigate to that end point with
ever-increasing levels of success. The information those users seek — about
government, products, news, or health information — is in demand and the
smart owners of that information are making it readily available. Whether one
needs to know how to get a dog license, a camcorder, a weather update, or a
recommendation on allergy medicines, the winners in the new-old-new economy
are those that make this information easily found. Of real note for
enterprises is the fact that as the expectations of customers rises for the
availability of a daily weather report, so too have the expectations of
finding out pertinent, concise, unambiguous information about their products
and services. An IT vendor that makes finding such information about six-,
seven-, and eight-figure products or implementations unavailable or
incomplete risks playing with fire, as that potential customer — with ever
increasing navigation skills — will have the wherewithal to find an alternative. IT vendor Web sites that are still largely seen as
marketing or publicity tools are doing their companies a disservice. Images
of preening CEOs, their schedules, or tributes to their brilliance or the
company’s unique place among the stars provide little or no help to potential
customers who, seeing this wall of self-promotion where they hoped to find a
path to what they sought, may simply go elsewhere. Increasingly, Internet
users know where they want to go, and how to get there. The enterprise that
impedes that effort does so at its own peril.
|
|
|
Embedding Intelligence: MontaVista
and Microsoft Look to the Future
By Charles King
MontaVista Software has announced its new Linux
Consumer Electronics Edition (CEE) 3.0 and operating system and cross
development for consumer electronics applications such as digital
televisions, set top boxes, automotive telematics, and mobile phones.
According to the company, CEE is tailored to the demands of highly
constrained mobile devices which require low power consumption, restricted memory,
and short boot and resume times. CEE will initially support Texas
Instruments’ OMAP 1510 and OMAP5910 applications processors, the IBM
PowerPC405LP and the associated Artic II/III Reference Board. MontaVista’s
current consumer electronics customer list includes Sony, Panasonic
(Matsushita), NEC, Toshiba America, Zintec Holding, and Araneo. The new Linux
CEE 3.0 will be available worldwide in the first quarter of 2003. In an
unrelated announcement, Microsoft’s Bill Gates said that the company will
help develop “smart” wristwatches with manufacturers Citizen, Fossil, and
Suunto that will reach the market later in 2003. The new watches will be
based on Microsoft’s Smart Personal Objects Technology (SPOT). SPOT enables
devices to interact wirelessly with owners’ PCs or DirectBand, a planned
nationwide WAN based on FM subcarrier technology and new radio protocols to
access news, weather information, sports scores, and stock closing prices. No
pricing information for the new watches was discussed.
These two announcements offer starkly different
approaches to how technology will enhance future consumer devices.
MontaVista’s development of Linux-based embedded operating systems and
development tools turned the rather staid and proprietary embedded systems
sector on its head. Well established players such as Wind River have had to
shift strategies to keep up, and relative newcomer Microsoft has seen its
plans to drive embedded Windows CE into set top boxes and other consumer
electronics gear hit a bump in the road. MontaVista has not fractured
relations between Redmond and consumer electronics vendors who use MontaVista
products (and have sworn to continue using Windows CE). However, the
attraction of CEE 3.0 is simple enough, since it allows tech-savvy vendors to
both escape sometimes onerous OS licensing fees and retain more autonomy in
product/market development. So does MontaVista Linux CEE 3.0 qualify as a big
deal? Shortly before Christmas Sony and Matsushita (both of whom are
investors in and customers of MontaVista) announced their intention to
develop a Linux-based OS for consumer electronics products that looks
remarkably like CEE 3.0. While this week’s announcement may qualify as
something of a second, commercial coming for CEE 3.0, we regard this wider distribution
as potentially disruptive for Wind River and Microsoft, and an opportunity
for developers and vendors to experiment with a solid, field-tested Open
Source OS that can support a wide range of consumer electronic devices.
While Linux may be making inroads in some parts of
the consumer electronics space, what are we to make of SPOT-enabled “smart”
watches? First, it is important to remember Microsoft’s long strategic
efforts to move outside of the box (the PC box, that
is). The company has pursued any number of efforts, from new operating
systems to productivity apps to enterprise business computing to set top
boxes to cable television investments to electronic gaming devices, for good
reason. When the PC market takes one of its cyclical dives, income from other
sources helps keep the ship afloat. So how do “smart” watches fit into this
scenario? To us, the watches are the smallest part of this new initiative.
While they might appeal to some gadget-obsessed individuals (the company
admits its target demographic is 15- to 35-year-olds), the real story here is
DirectBand. If it works as the interactive broadcast medium Microsoft plans,
the company could become the owner of a nationwide information network that
would support, enable, and enhance any number of new Microsoft-enabled or
supported devices. “If” is the operational issue here, since such a network
would be hugely complex and expensive to deploy. But if Microsoft succeeds,
it could offer the company a way into a much bigger “box” than PC makers ever
imagined.
|
|
|
Internet Sales Tax on Its Way?
By Jim Balderston
Michigan legislators are on the verge of approving a
measure that would allow the state to collect taxes on purchases made over
the Internet or through catalogues. The legislation would be modeled on an
ongoing effort to simplify sales tax codes on interstate purchases so that
states could collect those revenues. Thirty-four states have or are
considering a simplification of sales taxes, according to the Streamlined
Sales Tax Project. The goal of this project is to streamline taxes for all
interstate sales, not just those online. While more than thirty states have
begun the process of participating in the program, the agreement will not
take effect until at least ten participating states (representing 20% of the
nation’s population) have amended their tax laws to comply with the
Streamlined Sales Tax Project’s uniform definitions of tax rates for
interstate and online goods and services. Michigan
officials say they are getting pressure from local, in-state businesses to
amend sales tax regulations, so that out-of-state retailers do not have the
advantage of avoiding Michigan’s 6% state sales tax.
With states across the nation facing huge budget
deficits, it is not surprising to see some movement to end the tax moratorium
on Internet purchases. With Internet retailing still showing huge
year-to-year growth numbers — despite a miserable holiday shopping season for
their bricks and mortar cousins — the image of a frail, nascent industry in
need of protection is getting a bit frayed at the edges. With states
considering huge budget cuts and various revenue schemes, including casinos,
the momentum to do away with a special dispensation for the Internet
retailers seems largely irresistible. Before any online retailers get too
wound up about Michigan’s move, they should stop and note that the
Streamlined Sales Tax Project is nowhere near being implemented. Even with
the most enthusiastic support imaginable in various state legislatures, the
implementation of such a uniform tax code across state lines has substantial
and time consuming legislative processes to be negotiated before any real
change occurs. One only needs to remember that the federal government, not
the states, may regulate interstate commerce. Thus, the likelihood of
Michigan or any other state enforcing local sales tax rules against a New
Hampshire business remains slim, especially in light of the Supreme Court
1992 ruling regarding enforcement of state taxation policy against businesses
without an economic nexus in the state in question.
When one considers that many Republican-run state
legislators have aversions to any sorts of tax increases, it makes the
likelihood of enactment any time soon lesser still. Individual states like
Michigan may go ahead and declare taxes on Internet or catalogue goods but it
will do them little good if the merchant that sold the product to a Michigan
resident cannot be forced to collect the tax or submit information about the
transaction. Alone, states can do very little but kick up dust. But for the
sake of argument, what happens if the Streamlined Sales Tax process actually
accomplishes its goals and most of the states participate? For one thing, how
will they determine where an entity is located? Is it in Delaware, or
Montana, where there is no sales tax? Or are just its servers? Has it
incorporated in one of those states? Or does its warehouse reside there? How
states define the “location” of an Internet retailer will be a key element of
taxation of their business, and one can’t help but wonder if states like
Delaware, Montana, New Hampshire, and Oregon will get a host of new business
incorporations by Internet retailers seeking sales tax havens. Stay tuned.
|
|
|
PeopleSoft Announces Sales Incentive
Management Solution
By Myles Suer
PeopleSoft announced this week the general
availability of PeopleSoft Sales Incentive Management (SIM), which will be
tailored to various industry verticals. As of this announcement, the company
has released the product targeted at high tech and industrial organizations.
SIM will be sold as a part of PeopleSoft's Human Capital Management (HCM)
offering.
PeopleSoft’s action validates what has heretofore
been a niche occupied by Motiva, Callius, Incentive Systems, and Synygy.
PeopleSoft claims its offering is a full-featured solution, but a
feature-by-feature comparison reveals that several of the features found in
Motiva, which we believe has the most complete offering in this space, are
missing from PeopleSoft’s new package. However, it is important to recognize
that many of the offerings in this space have been developed to solve
different problems. For example, Incentive Management’s offering is focused
on how a firm manages its sales team, whereas Motiva is about collecting and
managing all the costs involved in selling a product. The question for
PeopleSoft and others in this niche is: Who is the real customer in the
enterprise and what does that customer need? From our perspective, the
compelling need is creating the ability for management to understand the real
costs involved in selling products — in particular the soft costs. These
costs can have a real effect on an organization’s success and profitability.
Additionally, the ability to know the real cost of the company’s sales
channels and partnerships can be a great help to an enterprise that is under
margin pressure. Given that these needs would normally be thought of as sales
and marketing in focus, we find it very interesting that PeopleSoft has
chosen to lead this niche from an HR rather than a CRM/e-sales perspective.
Nonetheless, it will be interesting to see whether the niche gains
significance due to PeopleSoft’s attention or becomes absorbed into the mass
of features in an ER offering.
|
