CNET también está disponible en español.

Ir a español

Don't show this again

Culture

Commentary: User demand spurs memory R&D

Intel and others are working on promising storage technology for portable devices, but it's unclear how soon it'll be available or how scalable or permanent it'll be.

    Research and development efforts by Intel and others to develop new mobile memory technologies are being spurred by the continuing increase in demand for storage capacity on handheld devices, driven by user appetite for more data storage, richer media and application functionality.

    The promise of inexpensive, ubiquitous, high-speed wireless connections that would enable this storage to be separated from the handheld device and centralized is unlikely to become a reality in the near future. Therefore, delivering device storage on the device itself will be increasingly important.

    Although the hype surrounding third-generation mobile networks suggested that pervasive devices could develop as always-on thin clients accessing data and applications from central servers, we do not believe such networks will be implemented until 2006-07 in North America (a year or two after Europe and Asia Pacific).

    See news story:
    Intel looks at leap in handheld memory
    No matter how fast networks become, if they not ubiquitously and continuously available, devices will need to be smart clients rather than thin clients. Smart clients require some level of application sophistication on the device stored locally and must then be able to synchronize with the network when it becomes available. This will drive the need for more resources on the client device, including memory capacity.

    The mobile market will also require lower-cost and lower-power mobile memory components. In the long term, the capacity cost per bit needs to drop substantially to make mobile computing useful and attractive. Because of battery constraints, a prime focus of development efforts like Intel's is finding storage solutions that require zero power to sustain the memory (unlike flash memory).

    As on desktops and servers, a major challenge in processing more information on mobile devices is the gap between the speed of processors and memory subsystems. There is a limit to how fast a disk can spin, resulting in much slower read/write compared to the processor speed. Therefore, the quest for the "holy grail" of mobile memory is partly a quest for a solid-state memory technology that can be employed as a cache between the processor and disk, or as a replacement for disk memory.

    However, as Intel and others make progress in developing storage alternatives that meet the clear market need for improved mobile memory, we expect increased diversification of memory types to optimize various characteristics such as speed, size, power consumption and cost. Some technologies are actually slower than current memory components--but are still attractive because they are cheaper and use less power.

    Although the storage technologies being developed by Intel and others hold great promise to extend power threshold for portable devices at a relatively low cost, it is not yet clear how many of them will be available how quickly, how scalable they will be or how permanent they will be--even if the memory is nonvolatile, the number of read/write instances could be finite. In the near term, we expect incremental gains in mobile storage capacity using existing technologies.

    Organizations should expect their mobile users to demand increased data capacity--and many users will demand more capacity than is currently available for mobile devices. Although manufacturers of mobile devices and components are striving to catch up with user demand for increased capacity, we do not expect major near-term increases in device capacity (or battery life). Accordingly, IT groups need to strive to meet user demands as far as possible within the constraints of the current incremental pace of performance improvement.

    Meta Group analysts Jack Gold, Dale Kutnick, Steve Kleynhans, David Cearley, Val Sribar and William Zachmann contributed to this article.

    Visit Metagroup.com for more analysis of key IT and e-business issues.

    Entire contents, Copyright ? 2001 Meta Group, Inc. All rights reserved.