'We have been seen as evangelists for this technology' - Martin Kidgell, Vice President

-Martin Kidgell, Vice President
and Managing Director National Semiconductor, Asia Pacific

More than a year back, National
Semiconductor acquired Cyrix and its PC processor business, launched the infamous socket-7
processor, took on the might of Intel and then with a swish, which left many gasping,
resigned the scene to this foe. Many nodded knowingly-"Yes, they got it all wrong.
They should never have challenged Intel".

But now the gameplan is clear.
National Semiconductor withdrew to concentrate on the embedded processor and information
appliance business. The appliances market promises to dwarf the PC or fat or thick client
in the not so distant future. Its affair with socket-7 was an out of place moment of
adultery. National has played its cards and the stakes have never been higher.
Evangelizing the thin client and the appliances market is now the only ticket to its
survival. While many will piggy-back on its efforts, including Intel, there is no turning
back for National.

DATAQUEST's Arun Shankar caught up
with National's Martin Kidgell, for an indepth discussion on their information appliance
strategy. Kidgell has been working with National for the last fourteen years in various
management positions in Sales and Marketing. He began his career with National in the
European division in 1984. Excerpts of the interview:

Your technology vision looks at
information appliances replacing PCs in terms of sheer volumes in a few years from now.
Would you agree that information appliances are essentially thin clients in terms of the
underlying technology?

A thin client is a subset of information appliances. It is really targeted at the office
environment and for business. You could argue that a Webpad is a thin client for the
consumer environment, but I think if we start talking about thin clients for consumers,
the consumer does not know what they are and their market does not know what it is. It is
best to stick to terms that the industry understands.

But information appliances would still be networked?

Yes, otherwise it does not make sense.

So how do you see PCs, thin clients and information appliances coexisting?

People buy computers today to send emails, chat and get on the web. Nobody buys them for
spreadsheets or for WP. I do not see the PC being obsolete. In fact, I see this huge
opportunity for a whole new category of products called information appliances (IA) to sit
underneath the PC. It would be usable by your parents and grandparents who don't want to
get used to a PC. We have had this vision for 3-4 years and we expected the market to take
off quicker than it has done. It is certainly taking off now.

Wyse, HP, DXL, Philips, AOL have announced IA products in the last six months based on
Media GX. People have actually, for the first time since we started evangelizing these new
products, started coming to us. We have been seen as the evangelist for this technology
and we have to make sure we can realize the revenue. We don't just want to be the guys who
talked about it and did not make the money. We are working with Microsoft, QNX, Planet
Web-all those people and the box builders.

Are we seeing consumer and PC vendors create their own vision of information
appliance?

People are coming at it from their own companies. The PC people are coming at it from-how
do they defend their PC turf. They see people like Wyse becoming very successful. There is
also the big need to replace the 3270 green screens, because most of the OS behind them
are not Y2K compliant or not very user-friendly. But people who have the 3270 terminals do
not want to get into PCs, because of the frightening expenses that go with it. In the US,
the average cost of maintaining a PC per year, in a network, is $6,000.

These devices can be connected via networks, wireless telephony or other methods.
Does this affect the design of the appliance processor?

Not a great deal. Only the functionality of the access devices will change. You have many
different connections-modem, ethernet, hub or RF. What changes is how you format the data.

What about the range of transmission for wireless networks?

We designed the Webpad to support three different types of standards. We see that as an
interim solution, till we get higher bandwidth standards. The three standards we have
today tend to be for inboard or in-house applications upto 100 meters. For the people who
want to put it on GSM, you have full GSM network support.

But applications are still not deployed on the GSM network?

All it means-to come back to your analogy-the Webpad is acting as a thin client for a
consumer. All the services are being provided by the server. GSM is just acting as the
connectivity mode. GSM today has a very low bandwidth, 14.4Kbps. Next year we get GPRF,
which is the next phase of GSM, which takes us up to 144Kbps. The year after we get GSMH,
which takes us upto 384Kbps of data. So by 2001, we have a very interesting bandwidth
available in GSM. And the GSM operators are starting to look for things that can exploit
that bandwidth. That is why they are talking to us.

With applications, internet and GSM merging in the same appliance, what will happen
to the service providers?

I believe service providers are going to become one. If you look at it-internet service
provider, telephony service provider, cable service provider, GSM service provider-I think
overtime we are going to start dropping the 'I', 'g' and the rest of it. So you have
service providers that provide you the bandwidth and content providers that provide you,
with whatever is on that. In some areas they will be the same. AOL is a bit of both. In
Hong Kong and Singapore we have two telcos, who tend to be a bit of both ISP and CSP.

What are the current limitations for system-on-a-chip design?

The limitations today are really the relevant tools to stitch it all together-in case you
are coming out with mixed mode tools for digital and analog. So we have to be working not
only with silicon manufacturers but also with tool manufacturers. The time for digital
versus analog testing is very different, with different methodologies.

We have not seen any efforts to embed memory-on-the-chip?

And we do not intend to. It is not cost effective to put memory on the chip. It just does
not make sense. You will see a certain amount of cache on that, but you have to leave the
memory out of it. It is a different process technology. And it is not going to happen in
the near future. Memory will remain external.

Will the nature of the OS be an important factor behind the large number of
appliances?

If you look at appliances there is no differentiation between the OS and the application
layer.

But is that the way we want to go?

Probably. If you look at appliances today-you cannot recognize the OS in the TV. The OS
should be invisible to the user. All you should have is to be presented with some sort of
graphical interface. And pull-down menus are not intuitive to anybody who has not used a
computer.

What is the future of Windows CE?

Windows CE certainly has a future. But it will be complemented by lots of other OSs. That
is why we are working with Microsoft and also with a number of others.

Will Windows CE be important in the thin-client market segment?

It depends from where the manufacturer is coming. If he is from a PC environment, he tends
to be comfortable with Microsoft. If he is coming from the consumer area, he will not
touch Microsoft -I am not going to let Microsoft take a big chunk of my revenue.

If you are coming from the communication area, like Samsung, Nokia, Ericssion-they
already have an OS in the cellphone. As a group they decided to go for Similar Epoch OS
from Psion. Again, they are not going to let Microsoft cream off anything from their
business.

What about application service providers and appliances?

This is my own personal view-I think that is something that IA will allow. If you want to
run a spreadsheet, why do you want to have that program on your device. You should be able
to run that off a server.

What is the global vendor scenario for appliance manufacturing?

We are seeing a huge concentration of manufacturers in Taiwan. The current PC vendors are
only marketing thin clients. We have to work with HP, Wyse, AOL, Phillips and the people
in Taiwan who make boxes.

What is the future of the appliances industry in India?

Realistically, India is not a major manufacturer of PCs and I do not see why it will be a
major manufacturer of IA devices. But I think there is an opportunity to add value.
Particularly in India and a few other countries in Asia, where they do not have a strong
manufacturing base, we are saying-do not try and duplicate what somebody else has done. Go
and innovate in the layer above that. There are significantly more margins in it. Less
capital investment and skill sets in India are more tailored for it.

In Australia, we believe it is going to be content that drives appliances because they
have innovative marketing.

How will analog and IA products affect your revenue realization in the future?

The thing that generates all the money for National today is our analog business. It
provides 70% of our revenue of $2 billion and provides the profits for investment in the
IA area. The IA business will account for 10-15% of revenue in the current fiscal year.
Going forward, we believe it is going to grow much faster than our core analog business.
We are not abandoning our analog business at all. It is a huge part of our company, but
more and more analog products are tailored to go into IA-type products.

We expect our analog business to grow close to the market rate, which is 8-12% a year.
We expect the IA market to grow significantly faster. We are expecting 100% per annum
growth from IA over the next few years, which is why we are investing so much money in it.