P2PNETWORKS AS SI PATHWAY? It never rains but it pours [Fwd: Economist: Sun stroke]

From: Michael M. Butler (butler@comp-lib.org)
Date: Thu Apr 04 2002 - 20:26:29 MST


Some sloppy goshwow along with more food for concerned thought....

-------- Original Message --------
Subject: Economist: Sun stroke

=20
Sun stroke
Mar 14th 2002=20

Sun Microsystems=8Bthe largest maker of servers for computer networks=8Bis
abou=
t
to turn its mantra, =B3The Network is the Computer=B2, inside out. Will the
new
vision relegate centralised servers to the scrap heap?

http://www.economist.com/science/tq/displayStory.cfm?story_id=3D1020705

HISTORY repeats itself, even in the computer business. Decades ago, Unix
programmers formulated ways of getting the most out of their computers by
tapping storage drives, memory and processing power from countless location=
s
around the network. That way, they hoped to squeeze out every bit of power
for their number crunching. In the process, they created a virtual
computer=8Balbeit a relatively slow one by today's standards=8Bwith its
innards
scattered about the office or campus.

The idea is fast coming back into vogue, as a host of high-tech firms start
unveiling distributed computing efforts of their own. All of them aim at
tying disparate computer systems together to create a system that behaves
like a single unit. This time, however, the goal is to create new
capabilities instead of simply borrowing resources.

Over the past year, IBM, Compaq and Hewlett-Packard have all unveiled
integrated distributed computing efforts, seeking to sell products and
services that take advantage of this emerging technology. Yet by far the
greatest convert to the new religion has been Sun Microsystems=8Blong the
stalwart of network computing with the mantra, =B3The Network is the
Computer=B2. But the way things are going, that slogan could soon be =B3The
Computer is the Network=B2.

In February, Sun outlined its vision for computing within data centres, the
highest performance computer market. The initiative, called N1, borrows fro=
m
all the distributed computing work going on within the company. At its most
basic, N1 is a network architecture that builds a computer out of countless
bits and pieces scattered around the network. It employs many of the same
protocols used on the Internet itself, and combines both hardware and
software to achieve its effect.

At the heart of it all is an effort to =B3virtualise=B2 computing systems,
turning servers and file storage systems into buckets of computing power
that can be connected like Lego blocks and managed just as easily. In doing
so, N1 allows computers and storage systems to be assembled into a much
larger system that can be treated as a single unit. As Sun's chief
technology officer, Greg Papadopoulos, notes, this was the same thinking
behind operating systems such as Unix, which was designed by some of Sun's
top executives in the 1970s to manage separate pieces of hardware as if the=
y
were all part of a single computer.

For Sun, the initiative emphasises the company's head-long dive into
distributed computing. Long a pet topic of Bill Joy, Sun's co-founder and
chief scientist, a host of distributed computing initiatives is being
pursued within the company. Last year, Mr Joy launched Project JXTA
(pronounced =B3Juxta=B2), a peer-to-peer computing effort which is seeking to
provide the wherewithal for connecting together countless devices ranging
from personal computers to refrigerators. Ultimately, Juxta resembles an
open-source operating system for peer-to-peer computing, providing a stack
of protocols that will allow all computers to find and connect to one
another without any form of centralised help.

In the same spirit in which N1 was launched in February, Mr Joy started
Juxta in order to create an industry standard for peer-to-peer
computing=8Band, in the process, encourage the development of new peer-to-pee=
r
applications. Juxta is seeking to build a peer-to-peer platform that is
available to any developer who wants to create applications on top of it.
Sun has enlisted a group of programmers from CollabNet=8Bone of the more
prominent communities of volunteer developers and hackers=8Bto help produce
the software.

Today, Juxta uses a stack of six protocols, based largely on the XML
(Extensible Mark-up Language), the programming language used to define a we=
b
page's layout and subject matter. In Juxta, the protocols let individual
computers (ie, peers) do various things, such as advertise their presence
and resources they possess; locate other peers; inform one another about th=
e
status of other peers; let various peers organise themselves into teams;
establish =B3pipelines=B2 throughout the peer community; and allow two or more
peers to borrow a pipeline so they can send private messages back and forth
securely. Underlying all the protocols is a security mechanism that prevent=
s
anyone from listening in.

The community aspect of Juxta has turned out to be more complicated than
expected. Unlike a corporate group of software developers, CollabNet brings
together a ragtag assortment of crack programmers who volunteer their
services for the prestige and personal satisfaction. That makes the effort
inherently political and far slower than a commercial one. =B3Grassroots
movements aren't easy to organise,=B2 says Mr Joy. Working together has been =
a
sobering experience for both sides. Nevertheless, Sun has found that
collaborating with the open-source community has helped the technology get
into lots of hands quickly, causing countless tinkerers to look at ways of
using Juxta in their own software programs.

But what do a project aimed at the highest performance end of computing and
one aimed at the lowest have in common? Quite simply, they both rely on the
fact that countless types, shapes and sizes of devices will soon be
communicating with each other. The concept of distributed computing=8Bie,
peer-to-peer networking=8Bhas gained popularity thanks to the explosion in
computing power, the arrival of broadband networking for the masses, and th=
e
widespread access to the Internet. Now that personal computers have more
muscle than the brawniest supercomputer of little more than a decade ago,
and even the simplest microprocessor-controlled widget has enough spare
processing clout to connect directly to other devices without outside help,
the time has come to start thinking about ditching the central organisation
that has ruled network computing from the beginning. Freed of such network
bureaucracy, computer systems actually become more robust, more reliable an=
d
perhaps even a good deal more secure.

Whether they know it or not, some 70m people have been enjoying this kind o=
f
network freedom for several years. All those who use an instant-messaging
service (eg, AOL Instant Messenger, Yahoo! Messenger and MSN Messenger) to
zip messages around the Internet are communicating directly with one anothe=
r
via a peer-to-peer network with almost no central control whatsoever.
Likewise, all those who used Napster, the now dormant music-sharing service=
,
were doing the same. The arrival of N1 and Juxta simply raises the
stakes=8Bmaking distributed networking that much more ubiquitous and easier t=
o
use.

=20



This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:13:15 MST