The Internet will eventually emerge as a "global networked computing
utility, replacing computing as we know it today," according to
PricewaterhouseCoopers' Technology Forecast 2002-2004, Volume 2:
Emerging Patterns of Internet Computing.
Internet computing of the future will result from the convergence of
several developments already underway, the report said, listing those
trends as Grid, utility, IP dial tone and ubiquitous computing.
"Grid computing enables large-scale aggregation and integration of
computing, storage, network and other information resources across
organizational boundaries, but its use today is limited to experimental
scientists analyzing large amounts of data," the report said. "In much
the same way the Internet began as an effort to share computing
resources among academic researchers before spreading to businesses and
then to consumers, Grid computing will eventually be used by
enterprises. More important, research into Grid computing will lead to
the creation of software that will solve critical problems companies now
face in building distributed inter-enterprise applications."
"Earlier this year, we predicted that Web services will take longer to
gain acceptance in inter-enterprise use than anticipated because of
unresolved problems with reliability, security and business models,"
said Eric Berg, managing director at PricewaterhouseCoopers' Global
Technology Centre and editor-in-chief of the Forecast. "Our latest
finding is that the development of Grid computing architectures,
protocols and middleware by the research community is likely to lead
directly to solutions that will overcome these obstacles and allow Web
services deployment to proceed."
The report highlights advances in computing and communications
technologies that are occurring in response to the increasing demands of
enterprise computing and use of the Internet. In addition to Grid
computing, several other trends foreshadow the development of Internet
computing, the Forecast said:
Computing as a utility -- an IT delivery mechanism in which
enterprises would no longer be responsible for operating their own
computing infrastructure any more than they operate their own
electricity generation plants or water reservoirs; instead, they would
access (and pay for) computing offered by a service provider.
IP dial tone -- "where the Internet becomes as ubiquitous as the
telephone network and as easy to access as picking up a phone." As this
evolves, the global telephone network will move from its current state -
islands of voice-over-Internet Protocol (IP) surrounded by an ocean of
legacy telephone system protocols - to the use of IP on an end-to-end
basis. In this progression, telecom carriers will lose the monopoly
they have on the creation of new telephone services, the report said.
Pervasive or ubiquitous computing -- "where IT is universally
available, but also becomes invisible because it is such a part of the
fabric of daily life. Of all the aspects of Internet computing, this is
the farthest from widespread availability, and its realization will
require advances in user interfaces; system software - including service
discovery, content adaptation, programming frameworks - and security;
and pervasive networks, including the creation of ad hoc networks
consisting of large numbers of sensor devices aware of their location
and capable of acting autonomously."
In the shorter term, the report said, three key trends characterize
major developments in computing and communications: the on-going
increase in the capabilities of commodity technologies; the increased
emphasis on scalable computing and communications infrastructures; and a
renewed emphasis on the flexibility, reliability and manageability of
those infrastructures.
"Flexibility, reliability and manageability are increasingly important
as enterprises must handle workloads that vary dramatically, ensure
availability of systems for use by employees, customers, and partners,
and operate IT infrastructures in a cost-effective fashion," the report
said. These requirements are accommodated by:
Resource management tools, previously available only on mainframes,
that are now available in distributed computing environments.
System vendors taking advantage of the falling cost of semiconductor
and storage components to build additional redundancy into their
systems.
New tools for system and network management that allow an enterprise
to model the impact of system downtime on business metrics such as
revenue or profitability, thereby being able to allocate resources so
the greatest investment in reliability is made in the areas that also
have the greatest impact on the business.
New architectures for data centers and Web applications that provide
enhanced flexibility and reliability.
According to Berg, "Enterprises today are finding it essential to
operate their computing environments in the most cost-effective manner
possible. During the boom, companies and service providers could just
throw resources at an IT problem until it went away. That's no longer
an acceptable strategy, so products that enhance flexibility,
manageability and reliability are at a premium in the marketplace."