[Fortune]Napster Is Clouding Grove's Crystal Ball

Date view Thread view Subject view Author view

From: Linda (joelinda1@home.com)
Date: Fri May 26 2000 - 11:27:21 PDT


[Some interesting thoughts from Andy Grove on Napster and its potential influence on network infrastructure.

--Linda]

                   http://www.fortune.com/fortune/technology/2000/05/29/eco.html

                   Napster Is Clouding Grove's Crystal Ball

                   Forget consumer broadband, says the Intel chairman in
                   a FORTUNE interview. To see the future, check out
                   what's happening at Indiana University.

                   We always like it when Andy Grove comes
                   to town. He spends a few minutes talking
                   about Intel, and then, with the prodding
                   of a single question, he's off on the
                   challenges of the entire tech industry.
                   Arguably, no one thinks about the big
                   picture as well as the 63-year-old
                   chairman of Intel. On a recent morning,
                   looking tanned and fit in a white shirt and
                   blue-and-gray tie and sporting a cheap
                   digital watch, Grove pondered the
                   devices that will connect us to the
                   future, called broadband a pipe dream
                   (literally) for consumers, and talked about
                   why Napster may mean Intel is about to
                   waste $80 million a year.

                   Do you think wireless devices or
                   devices other than PCs will become
                   the main points of entry into the
                   Internet?
                   Is an e-mail message or voice telephony
                   handled by packet switching an entry to
                   the Internet? If the answer to that is
                   yes, then my answer is yes.

                   Internet use is going to come in a big,
                   broad range of things, of which, I think,
                   only messaging--either voice or
                   textual--is likely to be a heavily formed
                   wireless business.

                   So are you down on broadband?
                   I have been making a very good living as
                   a prophet debunking consumer broadband
                   for five years, and I don't see any reason
                   to change.

                   Getting out to the consumer is a very
                   expensive undertaking. It's also a very
                   difficult technical undertaking. In the
                   United States, it is on the order of
                   a--and this is an illustrative
                   number--$100 billion infrastructure
                   investment. And while the tolerance of
                   the markets for investing $1 billion with
                   no foreseeable return has increased, the
                   markets have not reached the point of
                   tolerating investments of $100 billion that
                   have no foreseeable returns.

                   Twenty dollars a month gets you a basic
                   telephone connection to the Internet. To
                   pay for [broadband], you have to charge
                   something in excess of that. It's like a
                   first-class seat: It has to be more than a
                   few millimeters bigger than a coach seat.

                   There are plenty of businesses
                   desperate for bandwidth. You are
                   talking only about consumers, right?
                   Exactly. As all this is happening,
                   bandwidth is being deployed in very, very
                   major ways as backbone bandwidth and
                   as connections to businesses.

                   [Intel is] making a contribution to the
                   bandwidth economy. Recently we had a
                   strategic review organized by our CIO,
                   who just finished revamping our internal
                   network and wide-area network--a
                   project that we started in the mid-'90s.
                   It was a very expensive project. I forget
                   the number. It's a $100 million, $150
                   million kind of project. And the blueprints
                   [showed] this as a network that was
                   going to take us to the next century. And
                   it did. Just.

                   So now you're getting ready to pony
                   up just how much more money?
                   Eighty-million dollars a year for the next
                   five years. We are not getting another
                   application in there. We are just
                   expanding international capacity.

                   Do you see streaming video as
                   something to drive demand for more
                   powerful processors?
                   It will. But I have a question about how
                   mainstream that is going to be. Now,
                   having said that, let me argue in a
                   completely opposite direction. One way
                   around two problems we are talking
                   about--the broadband access problem
                   and delivering the higher-quality
                   experience of streaming video--is going
                   to come in a very dramatically different
                   network architecture than what we are
                   dealing with today. You get glimpses of it
                   in the Akamais of the world and even
                   better glimpses of it in the architecture
                   that's behind Napster.

                   The University of Indiana has become the
                   most high-profile deployment place for
                   that. The Napster traffic is on campus.
                   It's a reversion of a large portion of
                   traffic back to some modern equivalent of
                   local-area networking rather than
                   wide-area networking. It doesn't have to
                   be music or video. It can be corporate
                   data.

                   My worry about the building up of the
                   networking infrastructure that is costing
                   us $80 million a year is that we're going
                   to build that infrastructure just when the
                   usage pattern is going to a much more
                   peer-to-peer-based orientation. It shifts
                   the bandwidth requirement necessary to
                   go from us to the cloud [servers storing
                   data and applications] to bandwidth
                   necessary to go from my computer to his
                   computer, which may be in the cloud.
                   That may be the shortest way, or maybe
                   it's going to be a local-area network, or
                   maybe it's going to be some wireless
                   contraption. If the distances required for
                   these connections are reduced, then we
                   can solve the bandwidth problem without
                   solving it in a universal fashion.

                   There are some real trends of
                   decentralizing the cloud and bringing it
                   closer to the personal computer and even
                   involving the personal computer as nodes
                   in the network. This, of course, goes
                   back into the post-PC deal and throws all
                   that stuff [about the network computer]
                   out, because the device is too
                   underpowered.

                   So this is a wild card that can have a
                   profound effect on network build-out and
                   what has to play on the network.

                   feedback: ecompany@fortunemail.com

                   Vol. 141, No. 11
                   May 29, 2000


Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Fri May 26 2000 - 12:15:48 PDT