The recent slowdown in the PC market has revived old debates about the need for ever-greater processing power on client systems.
While one camp contends that people can never have enough processing power, an increasingly vocal faction contends that computing power outstrips the needs of customers and that further improvements are unnecessary. This debate is being fueled by the continued technological progress in developing faster processors.
We expect desktop systems of 10GHz or better by 2005--and more powerful systems after that.
In the consumer sector, we believe that increasingly more complex games and ever richer entertainment will drive processing demand upward. In the corporate market, we believe that a wide range of new uses will emerge during the next two to six years that will consume more processing power at the client. Although no single "killer application" will drive PCs to 3GHz, 5GHz or 10GHz, we expect a
wide range of new software and components that will, in aggregate, severely tax current PC resources and necessitate the growth in computing power.
However, because the increases in demand for processing power will be steady and incremental rather than the result of a sudden change in software or hardware features, they offer no short-term salve for PC sellers that have been struggling to find a way to jump-start the consumer and corporate PC markets. Likewise,
organizations responding to the current economic downturn by lengthening desktop and laptop life cycles beyond three years will face no near-term disadvantage from failing to immediately boost the power available on desktops.
To understand the innovations that will fuel the long-term need for more power, it is necessary to recognize the following major trends shaping computing in both the consumer and corporate markets:
• Greater collaboration, resulting from the combination of communicating and computing.
• The "always connected, always on" world, created by the Internet and wireless technology, driving the need for faster response times, security and availability.
• Digital media replacing all types of analog media.
• Increasingly global organizations struggling to foster worldwide collaboration and reduce linguistic and cultural differences.
• Decentralization of the work force (via telecommuting and so on).
• Greater prevalence of the "knowledge worker" model, encompassing all employee levels and job types.
• The need to reduce computing complexity, enabling people to derive maximum value from their systems.
Greater complexity demands higher capacity
The most obvious driver for new computing power has always been increasingly complex applications. Although most business computing has remained rather mundane (that is, pushing text and numbers around), more advanced computing techniques are starting to find their way into everyday (albeit specialized) usage--including computer simulations, rich 3D visualization tools, advanced learning systems and contextual adaptation.
We do not expect all people to have a need for these applications, but developers will inevitably begin to use these techniques in a wider range of applications, fueling the need for more power in an increasing number of client systems.
Administrative and maintenance tasks for PCs will also use more processing power as it becomes available. These include system "housekeeping" tasks (via tools that monitor, correct, update and optimize the system), reliability enhancements (such as system file protection and redundancy), the elimination of
performance "bottlenecks," and synchronization of information between devices using the PC as the central hub. Microsoft has already begun to implement some of these capabilities in Windows XP and Windows 2000 Professional, and more advanced features will continue to emerge.
Computer users will also benefit from improved information management capabilities to cope with the rising flood of information arriving at their desktops via e-mail, instant messages, voice mail, Web sites, and so on. New system capabilities will likely include enhanced e-mail management technologies, processing capacity to handle multiple streams of rich media and the metadata (such as XML tags) associated with content and interactions, and enhanced encryption and other security capabilities resident on the client system.
Between 2002 and 2006, companies will aggressively embed collaborative components into business applications, enabling shorter sales cycles, faster product development, increased transactions, improved partner and customer retention, and expedited problem resolution. Collaboration will extend to include streaming audio, voice, and (by 2004) video, driven by corporate communication and e-learning strategies. Collaboration will also expand to encompass real-time interaction and information exchange, making greater use of peer-to-peer interconnection to enable creation of ad hoc teams. All of these changes will demand more system resources.
From 2002 to 2004, we expect significant progress in bringing more natural interaction models to typical client systems. This will include speech, handwriting and visual input. On the output side, the current graphical interface will be extended to deal with the flood of new data types entering the market. We expect that, by 2005, PCs will devote as much as 90 percent of their processing power to user interaction. This will enable users to become less aware of the PC as a device and will enable it to play a greater role in daily activities. Even in cases where the PC is continually connected to the network with significant data and application logic on a server, we expect some application and much of the interface logic to exist on the client.
Want more? Pay (not that much) more
For business users, we believe the added value (that is, management cost savings, increased productivity, improved collaboration) of these and other PC enhancements, enabled by ongoing increases in processing power, will continue to justify the price difference ($150 to $300) between two similarly configured systems from a top-tier manufacturer--one with a higher-end processor and
another with a trailing-edge processor. On a typical business lease, this results in a difference of between $4 and $8 per month per user.
User organizations should draw a distinction between PC life cycles for systems already purchased vs. those for future purchases. During 2001, responding to sharp budget cutbacks, many organizations stretched their desktop life cycles by an additional year, delaying upgrades from Windows 98 to Windows 2000 (or Windows XP), and the same will hold true for 2002. Although this may be advisable for many organizations in the short term, we expect the long-term trend of increased processing power and functionality on the desktop to counter any pressure to permanently lengthen PC life cycles.
Although three-year upgrade cycles have long been the standard corporate best practice, some companies have always lived with longer desktop lifetimes--four or even five years. Whether organizations and consumers can live with a longer life cycle depends on several factors, including the needs of individual users. The question is how many within an organization would benefit from the capabilities of an updated PC, operating system and applications.
Regardless of how long their PC life cycle is, organizations should avoid fragmenting their user base between different operating systems and application versions. The more variations in the PC image that the IT organization must support, the more complex and expensive that support becomes. As the power and
capabilities of client systems continue to increase, organizations must continue to periodically refresh their PC fleets based on the best practices that have evolved during the past two decades.
Meta Group analysts Steve Kleynhans, Mike Gotta, Jack Gold, Val Sribar, Jeffrey Mann, Timothy Hickernell, David Folger, Thomas Murphy, Dale Kutnick, Chris Byrnes and John Brand contributed to this article.
Visit Metagroup.com for more analysis of key IT and e-business issues.
Entire contents, Copyright © 2002 Meta Group, Inc. All rights reserved.