SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Silicon Graphics, Inc. (SGI) -- Ignore unavailable to you. Want to Upgrade?


To: Jerry Whlan who wrote (3653)11/22/1997 11:49:00 PM
From: vincent bilotta  Respond to of 14451
 
the NYT has a relavent article in the Sunday business section: nytimes.com it covers Merrill's decisions to outsource it's network and the why's of a commitment to NT. as far as i know, SGI has never "worked with" a large financial institution in the stratigic sense they're talking about here. the way they do very well with AHPCRC or SNL or ILM. Sun has 10 years of first priority servitude in this market, and MSFT is offering enterprise customers developmental imput, and pampered service. what is the SGI plan for this market? is the O2000 what EDS or Anderson or AT&T recomend?
vincent

The No. 1 Customer: Sorry, It Isn't You

By SETH SCHIESEL

he information revolution arrived at Merrill Lynch on Feb. 7, 1995.

That day, two of Merrill's technology chiefs suggested to Edward L. Goldberg, the firm's
executive vice president for operations, that the $25 billion brokerage giant find an outside company to
run its communications network.

"We're a financial services company, not a telecommunications
company," explained Bruce Sieben, a first vice president who
helped to make the proposal, recalling his rationale. "We started
to ask if someone could handle the network for us so we could
focus on our business."

Sitting at a brown marble table in his office overlooking New
York harbor, Goldberg had a ready response: "Why didn't you
come to me sooner?"

Fifteen months later the company signed a $400 million contract
with the AT&T Corp. to manage Merrill's vast communications
system.

But that huge sum was just a drop in the bucket of dollars that
Merrill is spending on technology in the last half of the 90's. In
addition to the deal with AT&T, Merrill is pouring $825 million
over five years into upgrading its financial information systems
and about $300 million to address potential "year 2000"
problems in its computers. And those big-ticket items barely
dent the firm's annual technology outlays of about $1.5 billion.

Spending power like that explains why large multinational
corporations like Merrill Lynch have become the Holy Grail of
the information technology industry. By one estimate, 88
percent of the industry's $455 billion in annual revenue comes
from corporate customers rather than from consumers.

So the biggest battles in the technology world - over everything from computers to software to arcane
high-speed communications services - hinge on strategies for winning big business clients, not
everyday users.

As British Telecommunications, GTE and Worldcom vied to acquire
MCI Communications, for example, they were eyeing the two-thirds of
MCI's revenue that comes from high-margin business customers.

In its tug-of-war with Netscape Communications, Microsoft gives away
its Internet browser because the real money in cyberspace software is in
selling companies tools to conduct business on line.

And as C. Michael Armstrong takes over as chairman and chief
executive of AT&T, one of his highest strategic priorities is finding
strong partners overseas - all the better to appeal to multinational
customers.

"You could sign up 100,000 individual users," said Gerry Pape,
AT&T's general manager for the Merrill Lynch account, "and get as
much business as you could get from one large corporation."

The stakes are just as high for big companies as they choose partners to
guide them through the silicon thickets. For a company like Merrill
Lynch, the nation's largest financial services firm, advanced
communications technologies form the crucial infrastructure that makes
analysts, traders and money managers more productive.

Bad decisions can hand a huge advantage to a competitor or drive
customers to new technologies that bypass traditional brokerages
altogether.

So Merrill Lynch, as it tries to build networks to link its 54,200
employees in 870 locations around the world, has become an archetype
of the customers that are driving the evolution of the high-technology
industry. And at the same time, the firm is a microcosm of the
opportunities and pitfalls lurking in most every company's digital
future.

. Kelly Martin works in a plain building that bears no corporate logo, on the west side of
Manhattan. The walls are thick, the floors are thicker and the security is far more stringent than
at Merrill's headquarters at the World Financial Center downtown.

There, Martin holds down one of the most hectic jobs
in the digital world: managing technology for
Merrill's corporate and institutional client group. He
is responsible for the complex web of machines that
make possible Merrill's equity and fixed-income
trading operations, as well as its investment banking,
research and institutional sales. Martin's portfolio also
includes the company's internal risk, credit, treasury
and finance departments.

"We're trying to integrate Macintosh security in
Australia, we have an expansion plan in Brazil, we've
got changes in the UK markets and in NASDAQ,"
Martin said one recent morning, explaining what was
on his mind. "I can't staff all these projects."

Years ago, someone in Martin's job would evaluate
various brands of software and hardware, select a
group of vendors and count on his staff to make the
products work together. But these days, the number
and diversity of projects on Martin's agenda have forced him to turn to outsiders to help make those
decisions for him.

"Software companies want to sell software; I don't want to buy software," he said. "Technology
companies are interested in inventing new products all the time. When I started I got all these business
cards of people saying, 'I've got this product, that product, a new version of this.' I don't need
products. I need partners that can help me solve my problems."

So Martin hires companies like Andersen Consulting, Cambridge Technology Partners, Electronic Data
Systems and IBM - companies that he says can tightly define a goal, figure out what technology is
needed to achieve it and execute the project crisply.

Globally, the Yankee Group, a high-technology research firm in Boston, estimates that the market for
such technology problem-solving has grown from $61.4 billion in 1992 to $154.2 billion this year.
That expansion has prompted realignments in companies like IBM, which once focused on making
computers and programs, and Andersen, once known primarily as an accounting firm. Providing
technical services has become the fastest-growing part of both companies.

"The entire vendor-client relationship has shifted," Martin said. "There was a recognition on the part of
the technology vendors that selling Merrill more product won't help them."

But outsiders have not been able to handle all of Merrill's requests. The foundation of the company's
technology remains a web of IBM mainframe computers. Merrill tried to cut costs by turning
management of the machines over to their manufacturer a few years ago, but IBM said it could not run
the mainframes for less than Merrill was already spending.

Merrill told IBM, " 'You handle our entire data center,' " recalled Philip J. Villani, Merrill's chief
administrative officer and Goldberg's right-hand man. "And they said, 'No, we can't. You're too big.'
"

That leaves Martin making difficult choices. At the company where Michael Bloomberg first introduced
his groundbreaking market information terminals, where derivatives traders started buying Sun
Microsystems work stations in bulk in the late 1980s, Martin must constantly resist the seduction of
innovation for its own sake.

"There are definitely people who always want to work on the sexy new stuff, and we do experiment,"
he said. "But we don't introduce things just because it's the new thing."

These days, that means he is standing in the breach between his trusty mainframes and the trend that has
swept over many companies in which information and processing power are taken out of back offices
and installed on desktops.

Employees throughout the company are saying, " 'Give me the information; I'll just figure it out on my
spreadsheet,' " Martin said. "And you have to balance that against allowing the company to sustain
itself by being able to manage information centrally.

"I don't think there's an ultimate stage where we say, 'This is it,' " he said. "Things in technology are
always in constant motion. The hardest thing is figuring out what to change and what not to."

echnology is moving at an easier pace at the serene Somerset, N.J., technology campus for the
company's retail business, 45 miles southwest of Martin's nerve center.

In a climate as different from Martin's as a 100-share stock trade is from a $100 million corporate
buyout, Merrill's retail unit is about midway through an $825 million effort to revamp the computer
system used by the company's 14,000 retail brokers and 11,000 other employees around the country.

The challenge for the retail group's technology team has not been to handle a torrent of crises; the ways
that people save for retirement or a child's education do not change as abruptly as institutional capital
markets. Rather, the group's task is to make once-in-a-generation choices that will affect how the new
system's users do their jobs for years to come.

"We view technology as an investment, not an expense," said Howard P. Sorgen, Martin's counterpart
for Merrill's retail business. "Kelly has to contend with a very different set of variables than we do.
And it shows in the ways we manage technology."

The financial advisers at Merrill have been using the
same outdated computers for seven years. When
clients call for stock quotes or other market
information, the broker usually looks up the data on a
small, text-only screen using confusing keyboard
combinations.

Customers and brokers alike could get more
information far more easily by bypassing Merrill's
current system and using other information services
and the Internet.

So a few years ago the company decided to upgrade
its desktop computers to give its brokers access to
more sophisticated sorts of financial data.

Like most technology executives, Sorgen wanted to
adopt a single system throughout his operation, so
that problems could be dealt with more uniformly and
future upgrades would be smoother. And though,
unlike Martin, he was not looking for a partner to help
run the project from start to finish, he did want to use
software from a company that would be almost as invested in his development process as was Merrill
itself.

His choice came down to one that is dividing corporate purchasing and technology departments around
the globe: a network from one of the many vendors that use the tried-and-true Unix operating system,
which has dominated the corporate market for decades, versus taking a chance on Microsoft's new
high-end operating system, Windows NT.

Sorgen went with Microsoft, because he figured he wasn't taking much of a chance, after all. As
perhaps the most visible Windows NT adherent in America, Merrill, he calculated, would receive
Microsoft's highest level of attention.

"There's no question that Microsoft was willing to work closely with us," Sorgen said. "We're a big
customer of theirs."

For Microsoft, collaborating with Merrill makes sense, too. Sales to big corporate customers accounted
for about a third of Microsoft's $11.4 billion in revenue last year, according to Paul Stanton, director
for marketing at Microsoft's business customer unit, and that proportion is growing. Big customers like
Merrill provide more than sales. They provide credibility.

"The way most customers make decisions is, they look to their peers in the industry and see what they
are doing," Stanton said. "Getting someone like Merrill Lynch to adopt our product is very important
and influential for us, because they are so widely respected."

Indeed, technology vendors trying to introduce new products or elbow past competitors engage in a
game of cat and mouse. "I get 35 visits a week from technology companies, and they all tell me the
same lies," said Howard Anderson, managing director of the Yankee Group. "They all start with
'Merrill Lynch is my customer; Bear, Stearns; Du Pont; GE is my customer.'

"The idea is share of mind first, share of market second," he explained. "If I can capture the hearts and
minds of the most demanding of my users, the rest will follow along."

The big user, meanwhile, ends up with software tailored to its needs, without having to underwrite the
whole development process. Companies like Merrill work with developers like Microsoft before
software packages are even released, giving the customers a head start on the competition and the
developers critical feedback.

"They are a very committed early adopter of our products, and that obviously gives them more influence
with our plans," Stanton said of Merrill. "By them having a very close relationship with us, they wield
a lot of influence with us when we're making trade-offs on features for products."

For instance, when Merrill suggested ways that Microsoft could improve how Windows NT handles
complex communications among computers, Microsoft agreed. And when Merrill asked if Microsoft
could change how an electronic mail program sorted messages, Microsoft complied.

"This is a reciprocal relationship," said Tony Pizi, the Merrill executive who manages the company's
technical relationship with Microsoft. "They get a trial and we get responsiveness."

Large companies are vital clients for the computer and software industries. But in neither of those
sectors have corporate customers assumed such importance as they have in telecommunications.

Corporations, for example, have long had a choice of local telephone companies, while most consumers
still do not. The thirst to serve companies, meanwhile, has helped to drive the pace of
telecommunications merger activity to such a frenzy that $15 billion deals are now called medium-sized.

Hence a company like Worldcom could grow from nothing in the early 1980s to become the No. 4
long-distance telephone company by serving medium-sized business customers almost exclusively.
Now it has an agreement to use its high-flying stock to buy MCI for $36.5 billion, and Wall Street
analysts say that combination would allow Worldcom to compete more effectively for the biggest-ticket
business accounts.

till, as recently as last year, Merrill Lynch ran its own communications network. "The carriers
were not ready to take on the responsibility of managing a network the size and scope of Merrill
Lynch's," said Sieben, the first vice president, who has helped to oversee Merrill's
communications system.

The company leased bare-bones capacity from MCI and maintained a staff of almost 100 to keep
conversations and its huge data stream moving across the country.

But as Merrill ballooned along with the bull market of the 90s - and as the complexity of its
communications needs increased - the company finally decided to turn over the job of running its
network to an outside contractor.

AT&T, MCI and the Sprint Corp., the nation's top three long-distance companies, all bid for the Merrill
account. But Goldberg, the executive responsible for most aspects of Merrill's operations, from
communications to sanitation, said the firm's choice was not difficult. "We felt that we wanted the
expertise, the know-how and the discipline of AT&T," he said.

Those are not the sorts of words people in the telecommunications industry have been using to describe
AT&T recently, as the company suffered embarrassing turmoil in its executive suites and the collapse of
its merger talks with SBC Communications, formerly Southwestern Bell.

But Goldberg's description may point to AT&T's least-appreciated asset: its simple ability to cope with
huge projects like Merrill Lynch's.

It was not just AT&T's ability to meet Merrill's rate demands that won it the business.

"There might have been some companies out there, a bolt from the blue, that say, 'We can run your
network for 2.9 cents a minute and I can do everything you want,' " said Sieben, using a hypothetical
number. But promises are one matter, and performance is another.

"It's the easiest thing in the world to open up a telecommunications business," he said. "We did it 10
years ago. You buy some switches. You buy some remote access equipment. But the difference
between AT&T and XYZ is that AT&T has a service and support engine and a service and support
culture - and XYZ doesn't."

After AT&T and Merrill Lynch signed their contract in May 1996, about 80 Merrill employees became
employees of AT&T, even though they still work in Merrill's electronic fortress on the west side of
Manhattan. AT&T is paid for the performance and reliability of the network, on a sliding scale that
becomes more demanding over time.

Handing off the job has freed people like Martin, on the institutional side, and Sorgen, on the retail
side, to concentrate more on the substance of Merrill's business and less on technical details.

"Across the whole company you see the sorts of projects people are working on have shifted from big
infrastructure and network projects to applications sorts of projects," Sieben said.

And Merrill Lynch is not finished picking winners. AT&T's deal with the brokerage firm covers only
domestic operations; Merrill is getting ready to farm out its international network, as well - a contract
that could be even more lucrative than AT&T's United States agreement.

The phone companies are bound to line up for the job, just as they hanker after the telecommunications
business of the rest of the Fortune 500.

"You go to a Sprint large-user meeting, an MCI large-user meeting, an AT&T meeting, you see Exxon,
General Motors, EDS," said Anderson of the Yankee Group. "These guys are the business. They drive
the networks. They drive the innovation. What they do you follow."



To: Jerry Whlan who wrote (3653)11/23/1997 1:52:00 AM
From: rumboman  Respond to of 14451
 
To Jerry:
The following material is from Sun's press release about its
benchmarks:

The single system SPEC95 results were performed on a 64 CPU Starfire
server running the highly robust Solaris 2.5.1 operating environment and
configured with 4 megabytes (MB) E-cache. The server configuration
included 16 gigabytes (GB) memory and 189 GB of Sun(TM) RSM Array(TM)
2000 storage. The Starfire server achieved 4289 SPECint_rate_base95 and
4945 SPECint_rate95 integer processing performance. The same system,
configured with 18 GB of Sun RSM Array 2000 storage, measured 5417
SPECfp_rate_base95 and 6013 SPECfp_rate95 floating-point processing
performance. The SPECint_rate95 results are 186% faster than the nearest
competitive results, by SGI'S Origin 2000, and demonstrate the Starfire
system's ability to achieve the enormous capacity throughput required
for floating-point and integer intensive applications.

I may be missing something but it appears that Sun is comparing
its 64 cpu machine against SGI's 32 cpu machine while implying
that the comparison is to an equivalent SGI machine. I do not know
what exactly is standard operating procedure in computer vendor
business practices; however, in its next press release, I expect
Sun to compare its high end machines to my little laptop.



To: Jerry Whlan who wrote (3653)11/24/1997 9:24:00 AM
From: Alexis Cousein  Read Replies (1) | Respond to of 14451
 
>This don't show so well on the SPEC numbers, but if you compare
>TPC-C numbers (and probably TPC-D, SGI's lack of publishing leads one
>to think their numbers are poor)

Don't bet on it. We'll see when they're published :].

>Sun is a lot more efficient than SGI.

They're bigger, which accounts for the fact they can release TPC numbers faster (the configs used in these tests are hardly trivial to assemble). SGI's working on that...