Compare and Contrast: What a
Difference a Year Makes
EMC has reported Q4 03 and overall 2003 financial results.
Total consolidated revenue for Q4 03 was $1.86 billion, 25% higher than Q4 02.
Net income for Q4 03 was $220 million compared with a net loss of $64 million a
year ago. Total consolidated revenue for 2003 was $6.24 billion, up 15% over
2002. Net income for 2003 was $496 million, compared with a net loss of $119
million the previous year. Business segments including Symmetrix, CLARiiON,
Centera, Celerra, core software and services, and connectivity revenues
experienced double-digit revenue growth from Q3 03 to Q4 03. The company
credited its broadened product portfolio, focus on information lifecycle
management (ILM), and an improving global economy for its performance during
2003. In a separate announcement, Intel announced Q4 03 revenue of $8.74
billion, up 22% over Q4 02. Q4 03 revenue set a new record, eclipsing that of
the previous high set in Q3 00. Net income for Q4 03 was $2.2 billion, up 107%
over Q4 02. For the year, revenue was $30.1 billion, up 13% over 2002. Net
income was $5.6 billion, up 81% over the previous year. For the year, the
company paid cash dividends of $524 million, and used $4.0 billion in cash to
repurchase approximately 176 million shares of common stock. The company noted
that IA microprocessor units set a record, the average selling price was
slightly higher, chipsets were at record levels, motherboards set a record, and
Ethernet product units were at record levels. In yet another announcement, Sun
announced Q2 FY 04 revenue of $2.89 billion, up 4% over Q1 FY 04, but down 1%
over Q3 FY 03. The company indicated its Q1 to Q2 revenue growth was the
highest since 1998. Net loss for the quarter was $125 million compared with
$2.28 billion one year ago. The company highlighted its release of twenty new
products at its SunNetworkSM conference in Berlin and
an alliance with AMD as significant achievements for the quarter.
For many in the IT industry 2003 will be remembered as the
year that the market finally began to turn around and enterprises actually
began to loosen their purse strings. Intel achieved record shipments in their
venerable IA CPUs as well as their supporting chipset, motherboard, and
Ethernet products. These accomplishments were also marked by impressive
financial performance, especially notable in what have been rocky times for any
vendor’s pricing power. EMC continued down the path of change, apparently for
the better, as it has gone from being a money loser to a moneymaker. The high
growth rates for its products is impressive, but even more so is that this
growth in some cases is coming from new products in an emerging market. Sun has
also seen fortunes change for the better. The hemorrhaging of cash it
experienced a year ago is down to that of a bad cut, but is clearly no longer
life threatening. Sun too announced new products and some change in direction
in its quest to recapture its once unassailable position in the marketplace.
So overall things are getting better, right? Well, yes and no. Intel, EMC, and Sun are each major players in the IT space, yet while we see some similarities we believe these players illustrate different positions in the market, each fraught with its own set of risks and opportunities. Intel has clearly maintained its enviable position as the 32-bit commodity-computing powerhouse, complete with both market dominating and ancillary products. However, in its quest to grow new a market, requiring a 64-bit solution in the form of Itanium, the company continues to fall short of its aspirations. EMC has also taken a lesson from the commodity-computing textbook and began to shift its revenue dependence away from premium-priced hardware to a new market in services and most importantly value-added, standards-based software attacking a new opportunity it has dubbed information lifecycle management (ILM). The challenge of course is cultivating the ILM opportunity, which will not be an automatic slam-dunk. Sun has staunched the bleeding, mostly through a significant reduction in the size of its organization, and is also offering new products. However, the company remains fixated on its “my way or the highway” value proposition. While Sun loyalists remain vociferously supportive, others are finding the siren song of other vendors rather enticing. There are fundamental differences between the market positions typified by these three players. Intel is a stable giant with undeniable past success that is failing to move forward in large part due to its inability to produce a seamless transition and growth path for its customer base. Sun is a shrinking giant that seems only to gaze in the section of the marketplace residing near its navel, but it too has failed to produce a comprehensible, seamless transition and growth path for its customer base, let alone new buyers. In contrast, EMC, no small entity itself, has taken the bull by the horns and put in place a path to expand its markets and create seamless transition and growth for its customers. While it faces a significant challenge in making this come true, EMC’s actions smack of a logically crafted, strategic plan — which is in sharp contrast to the other camps.
Bringing Web Services to Grid, and
Vice Versa
Akamai, The Globus
Alliance, HP, IBM, Sonic Software, and TIBCO have proposed new Web services
specifications designed to integrate Grid and Web services standards. The new
WS-Notification and WS-Resource Framework represent the first time a common,
standards-based infrastructure will be available for sharing across business
applications, Grid resources, and systems management. The new specifications
will provide a scalable pub/sub messaging model and the ability to model stateful resources including physical entities (such as
servers) to logical constructs (such as business agreements and contracts).
Access to these resources enables customers to automate solutions just in time
procurement with multiple suppliers, systems outage detection and recovery, and
Grid-based workload balancing. WS-Notification can be configured to
automatically trigger an action in the IT infrastructure, such as notifying a
supplier to replenish inventory once current inventory drops to a set level. The
WS-Resource Framework includes WS-Resource Properties, which defines how data
associated with a stateful resource can be queried
and changed, and WS-Resource Lifetime, which allows the user to specify the
period during which a resource definition is valid. According to the
proponents, the new specifications will help enterprise customers lower costs,
speed deployment, and enable integration across and outside of the enterprise.
Spending a couple of hours with the hit film Seabiscuit allows one to bathe luxuriously in the rosy
glory of professional horse racing, but rose-tinted glasses cloud the
actualities of the Sport of Kings. Overall, it takes thousands of hours of
dedication, patience, intensive training, and stable mucking to get a
thoroughbred to the winning post on time. What does this have to do with these
new Grid and Web Services standards? In a word: everything. Any successful leap
forward requires more than a little creative speech writing. Visionary language
is what gets people’s interest up and blood running, whether it comes in the
form of inspiring election year rhetoric or business technology strategizing.
In the cases of grid computing and Web services, both efforts have taken up
enough newsprint and ink to fell a rainforest and darken the Amazon, and while
both efforts boast enthusiasts across the IT industry, they have generally
created more light than heat for end users chilled by continually increasing IT
complexity.
With that in mind, do the new WS-Notification and WS-Resource Framework specifications have a chance to improve things? Perhaps so. At the end of the day, the success of grid computing and Web service solutions will not depend on visionary rhetoric but by the business value their use provides customers. In the case of WS-Notification and WS-Resource Framework, that value will be found in the ability to increase business efficiency by better managing IT infrastructure, automating business agreements and contracts processes, and setting specific rules and criteria for controlling those processes. What does all this really mean? In creating effective specifications such as WS-Notification and WS-Resource Framework, the Globus Project and its partners are performing a good bit of proactive IT stable mucking for future grid and Web service customers. By putting in the time and effort now to, proponents of these specifications hope to help their clients beat the field and cross the finish line in record time. By doing so, they should not only earn respect, appreciation, and future sales opportunities from their clients, but will also help drive the vision of fully integrated, automated, and interdependent IT infrastructure across the greater market.
Intel Outlines Broadband Wireless Vision
Intel articulated its plan to work with the industry to
dramatically drive down the cost and increase the availability of broadband
wireless at the Wireless Communications Association annual symposium. Sean
Maloney, Intel EVP and GM of the Intel Communications Group, stated that
Intel’s efforts in 802.11 (WLAN) and 802.16 Wireless Metropolitan Area
Networking (WMAN) that would help attract the next 5 billion users to the
Internet, particularly those in emerging markets such as China, India, and
Latin America. The vision outlined by Intel includes delivery of
standards-based silicon for both WLAN networking and cost-effective
and interoperable 802.16 WMAN hardware. The 802.16 silicon — that would
be certified by the WiMAX Forum that oversees the
compatibility and interoperability of 802.16 technologies — would be developed
and deployed by a growing ecosystem of wireless equipment makers and service
providers. Intel envisions a three-phased deployment that will begin with fixed
outdoor antenna installations, rapidly progressing to indoor antenna
installations, and in the third phase, WiMAX
certified hardware would be available in portable solutions for users who want
to roam within or between service areas. WiMAX
certified systems would also be used to connect 802.11 hot spots and
enterprises to the Internet. Maloney announced that Intel is working with
telecom companies, including Airspan Networks, Alvarion,
Aperto Networks, and Redline, to
develop and deploy WiMAX certified 802.16 equipment
based on Intel silicon. Siemens Mobile and Proxim are
also separately in discussion with Intel on areas of collaboration for the WiMAX technology market. Intel's 802.16
silicon is scheduled to be introduced in the second half of 2004.
That dreaded last mile: the object and obstacle of much
pain and affliction to service provider and customer alike. The reasoning goes
that if the cost and obstacles inhibiting those last mile(s) could be driven
down substantially, then users aplenty would flock to the Internet, and other
digital services provided over these connections. Combined with the freedom of
wireless connectivity, WiMAX certified 802.16 would
seem to be the cat’s meow in last-mile digital service delivery. Outlying
suburban areas would finally be served, as would rural locations, and even
those in underdeveloped regions bereft of most any digital communications
infrastructure. For the billion or so people who find themselves fortunate
enough to live in an industrialized part of the world, overcoming the last mile
through wireless would seem feasible, and perhaps even desirable. One of the
great ironies of many broadband solutions is that they are available in densely
populated urban centers, which may be the workplace for many in the digital
literati, but is not representative of the suburban lifestyle commensurate with
the digital elite. So in the industrial world, 802.16 could provide a desired
service in areas where existing solutions fall short. The enhanced throughput
of 802.16 could also overcome the severe limitations experienced by past WMAN
services such as Ricochet.
Where we disagree with Intel’s vision is in its ability to attract the next 5 billion users to the Internet and/or digital services. In a world where three-quarters of the population lives in substandard housing, deals with diseases long eradicated in the West, lacks running potable water, and scrapes by at a subsistence level of food (on a good day), having wireless access to the Internet is as helpful as providing air conditioning to an Eskimo in the winter. Although many would like to view all technology as the great savior that can bring an end to poverty and blight, the simple reality is that most living outside the industrialized world have barely attained the Physiological Level of Mazlow’s hierarchy of needs. Until such time that the world can provide for the basic needs of the majority of the population, we believe the uptake of the Internet by the next 5 billion will be significantly stymied, regardless of how cost-effectively Intel and its buddies can manufacture 802.16 technology and how successful it will be within the domain of those capable of reading this Roundup.
Sun Purchases Nauticus Networks and High-Performance Content Switches
Sun Microsystems has announced plans to acquire Nauticus Networks in a cash transaction for an undisclosed
amount. Nautilus Networks is a privately held firm that makes high-performance
content switches. The switches handle SSL traffic, security, virtualization,
and load balancing. Sun believes that Nauticus,
destined to become part of Sun’s Volume Services Products organization, will be
important for the convergence of compute and network services.
After years of focusing Solaris and SPARC research on
high-end big-muscle systems, Sun has recently spent some time putting legs to
their volume story and getting that end of the family moving again with
specific actions like acquiring (and ultimately killing) Cobalt, and forming an
alliance with AMD. For HP, the way to deal with growing the volume server end
of the business was to purchase Compaq, the volume leader, and play
economies-of-scale and growth-through-downsizing. The two strategies may work
for limited time, since with their installed base HP can continue to drive
sales of volume products and then shift customers into their Adaptive
Enterprise computing infrastructure. IBM had always been rather coy about the
volume space, but recently they launched aggressive European marketing
campaigns challenging Dell on pricing and established industry leadership with
both the x450 for high-end Wintel/Lintel applications, and the BladeCenter,
which now runs on POWER as well as x86 architectures. For Sun, the challenge
has been a bit different. Despite HP’s and IBM’s UNIX and proprietary systems
focus, they have both always had a leg in the Intel+Microsoft
game. Sun on the other hand has not only distanced itself from those two giants
but has maintained an adversarial position, particularly against Microsoft.
They have dithered back and forth about porting Solaris x86 products and their
ongoing commitment has been unclear. As a result, their customer base for those
products was shaky, thus weakening executive resolve, and the vicious spiral
continued. The recent rise of Linux gave Sun another chance to issue confusing
messaging, but after a couple of hapless volleys, they finally seemed to have
figured out what their strategy is via hardware and OS in the volume space and
appear to be sticking to it.
All of which leads to the purchase of Nauticus. While HP and IBM have stayed relatively distant from the networking side, Sun has maintained for years that the network is the computer. This reasoning has led them to their purchase of Nauticus and its content switch. The technology can be leveraged in their blade servers, and across their N1 plans in order to grow their scale out capabilities. To some degree, Sun will have to play catch-up with IBM in the blade space. On the other hand, Sun is comfortable in the networking space, as much of their client base is telecom, so they may be able to take advantage of their familiarity with the inherent architectural requirements and limitations on some fronts. At the same time, most of N1 has been created by purchases of component parts from other companies and bringing staff from many small companies together. While integrating disparate products is not unusual in the marketplace, most companies reserve their in-house development muscle for products of strategic importance. Using bolt-together techniques of development for N1 says a great deal about Sun’s strategic view of this capability. At least in this case, Sun has chosen to own rather than rent the technology. Nevertheless, Sun still runs the risk of creating a camel — which everyone knows is a horse designed by a committee — rather than a thoroughbred.