From semiconductors to microprocessors and then to a full-fledged IT company,
Intel has been re-inventing itself almost every 10 years. It is now shaping up
as a company that will provide building blocks to the Internet economy. Two of
the company’s key officials–Christian Morales, vice president, sales and
marketing group, and GM, Asia-Pacific operations, and Avtar Saini, director,
south Asian operations–talk about Intel’s strategy to take on the challenges
of the e-economy, in a conversation with DATAQUEST. Excerpts:
Today, many companies are re-positioning themselves as e-enterprises.
While the IT industry has been going overboard with the concept, what are Intel’s
initiatives?
Well, back in 1993 when it was predicted that IT would be the largest industry
in this world by the turn of the century, few people believed it while many
others questioned it. Our mission at that time was to be a building block
supplier to the computing industry, which was operating as a vertical industry
with 5-6 manufacturers across the world supplying thousands of computers. Today,
however, we are in a horizontal industry, where you no longer have a couple of
computer manufacturers who have their own operating systems, their CPUs, their
applications and also channels. It is now an open industry where we have the OS
developers, the CPU manufacturers, application developers, different
distribution channels–whether through the Internet or dealers and resellers.
It’s now a market for service players. Also, this industry has moved from
handling a couple of server-end computers to handling millions of computers.
We see today two major phenomena in the world. One, the wireless and the
computing industry are kind of merging. Two, we see the emergence of two major
devices–the extremely thin PCs (laptops) and mobile phones with browsing
capabilities–that would be used for accessing the Internet. While today we
have around 300 million laptops and around 200 million-plus mobile phones, I
expect these numbers to grow exponentially in 3-5 years’ time.
Keeping in view these changes, two years ago we announced Internet exchange
architecture (IXA) with the objective of providing building blocks to the
networking manufacturers in the world, including Intel. It was also an attempt
to standardize the industry along open standards.
Coming back to the example of the computer industry, if we standardize the
e-business implementation, in ten years it will become a $10 trillion industry–10
times the volume of business today. But it’s going to take some major steps,
like planning standards along the software building blocks so that it becomes
much easier to implement e-business solutions.
Also, once you have open standards, then there are a lot of companies who
just have to write to that standard. They don’t have to work for another
company’s proprietary. What this does, is help promote innovation.
How much of the positioning that is happening worldwide is fundamental,
practical change and how much of it is marcom?
Three years ago, when Craig Barrett became the CEO of Intel, he changed the
mission of the company: to convert it into a building block supplier of Internet
economy. Accordingly, he restructured the company. We are focusing on four major
building blocks of the new economy, which are published and open. First is the
client focus that allows us to look into the devices. Then is the technology for
building blocks for the server. The third one is networking and communication.
And the fourth building block is services like datawarehousing, Web hosting,
application hosting and also data centers.
We have also been acquiring some companies working towards the convergence of
voice and data over the Net. We are firm believers that the telecom industry is
also going to move from a vertical to a horizontal model.
When you are a company like Intel, you do a few things: you consider your
environment, you consider your strategies, and you deploy your resources
accordingly. Yes, you are right. In the starting phase it is basically marcom.
It is repositioning of the company in the Internet era. But then after you have
taken this first step, the whole thing changes. We are not looking at the world
from the computing industry point of view; we are looking at it from the
Internet perspective. Naturally, then the whole thing changes–from the kind of
product you plan to develop to the type of services you want to provide. All of
these become Internet-centric. Accordingly, we have re-deployed our resources.
The restructuring is along the client-server platform, networking and
communication, and services. Has the company divided its resources into
different groups associated with each of these blocks?
Yes, different business units are totally independent and have their own
objectives and resources.
What are these units called?
First, the Intel architecture group that takes care of desktops on the one hand
and servers on the other. These are the CPUs, chipsets, motherboards and other
such products. Then the networking group that takes care of communications and
networking. Next Intel’s online services, which is about data hosting and
application hosting services. And there is a group called wireless and
communication component. This group works on technology needed for wireless
systems. In networking, we have two business groups. One is the component part
of it where we provide the critical component to companies like Cisco. The other
group provides solutions at the systems level like switches and cache appliances
for the Internet. So there are five groups, some of which existed before the
restructuring while most of them were formed after it.
Intel started in 1968 and it has since been re-inventing itself every ten
years. It started as a company in memory chips. Ten years into its life, it
became a microprocessor company. Ten years later, it converted itself into a
computer company, and now it is shaping up as a company that will provide
building blocks to the Internet economy. This is how we view ourselves, every
ten years we expand our focus to look at the next big growth. Traditionally,
Intel as a company has been growing at 20—25 per cent.
The client area is huge for desktops, mobile devices and notebooks. Apart
from desktops, Intel has been traditionally weak in the other areas. How are you
addressing this?
We have different teams working in each of the different segments. Again, these
teams are further categorized to work on the value PC, the performance PC and
the workstation. We also have teams working on different categories of mobile
devices.
Battery life is such a major issue with mobile devices but it seems Intel
has not been able to provide alternatives for mobile computing devices that need
to be sufficiently low-powered. Why?
Contrary to belief, Intel was the first company to take initiatives towards
reducing power consumption in such devices almost eight years ago. Only
recently, we launched a chipset for such devices where you don’t have to
compromise on the power, speed or the battery life. We are also working on a
project to develop chips that will consume much less power than today. Also, if
you take a look at today’s market, there are P3-based mobile devices with
battery life of more than five hours.
From the technology point-of-view, the CPU consumes only 10 per cent of the
power in a mobile device. The rest of the power is consumed by other components–the
drives and other plug-ins. Also, a lot depends on the efficiency of the battery.
So, we at Intel are not only working towards developing more power-efficient
CPUs, we are also taking initiatives in other areas just to ensure that the
battery lasts longer.
If the CPU takes up only 10 per cent of the power, one way to make the
battery last longer is to integrate a lot more components in the CPU. Intel was
exploring this possibility aggressively under Vinod Dham. Why was the project
shelved?
As far as the integration project is concerned, it had some major issues. When
you integrate products or applications, you are adding a step to innovation; you
have all the best pieces and you try integrating them. What this means is that
all these pieces of the computer got to be going lock-stock. But this doesn’t
happen in today’s world. So by default, an integrated product–the CPU in
this case–would always remain a generation behind the latest technology. From
the market point of view, this is not the ideal situation. We did spend millions
on the project but decided to shelve it because we realized that PC on a chip
couldn’t become a reality.
The other reason why it cannot be a reality is that the process involves
taking certain low value components and integrating them with a very high value
and complex device. While such integration would make the new device more
complex, it also does not make much of an economic sense.
The trade-off is always the compromise on performance, modularity and
scalability. Also, if you look at the corporate environment, they want platforms
that are stable and the highest possible speed. This is the priority for them
and in order to be able to cope with these demands, one needs to compromise
somewhere else.
So, how would you describe the processor of the future in terms of its
architecture, the complex tasks it would be able to handle, the size and the
cost?
As technology becomes superior, we are sure to be able to provide better
performance at similar costs. That aside, we have been steadily growing as per
Moore’s Law and in the foreseeable future we don’t see much deviation from
this pattern. From the functionality perspective, we keep on adding new features
to be able to run new applications that are constantly being introduced. When it
comes to the bandwidth of the processors, we are coming in with the Itanium
processors. From the flat form perspective, improvement in the IO and LAN
connections would be made.
For the next one or two decades, Moore’s Law will continue to be
applicable. We also see the present day segmentation continuing. In the desktop
arena, we will continue to have the performance and the value desktops. While in
the notebooks, it will be the full-size, the ultra-light and the slim-size, at
the server end it would be the front-end, the back-end and the industrial
servers.
Shubendhu Parth in New Delhi