Where do your data live?

By Peter Glaskowsky, photograph by David Toerge -- Electronic Business, 9/1/2000

The ongoing proliferation of personal electronics has an inconvenient corollary-a proliferation of places where information is created and stored. Having a variety of ways to create data is good, but being obliged to keep track of data in many different locations is bad.

What we need here is a little more structure-an effective strategy for dealing with data that exist in many places at once

Take phone numbers, for example. I have phone numbers in my watch, cell phone, PDA, laptop, desktops at home and work, and the office server. The cell phone is the most convenient input and storage device for phone numbers, but it has limited capacity and can't store essential context like street address, employer name and job title. My Newton has enough flash memory to hold a complete copy of a company contacts list with many thousands of names-but Apple never made the synchronization function work very well, so it gets out of date. By default, the office server holds the live version of my contact information, but I can't get to the server except when I'm at my desk.

How about digital photos? I create images with a digital camera, a digital camcorder and a scanner hooked up to my home desktop, and I get still more images in e-mail and off the Web. I have Compact Flash and Memory Stick adapters that let me download images to my laptop while traveling, thus creating yet another storage location for digital images. There's no way to tell if any specific image has been copied to another device and is safe to delete, and sorting through all these different images and putting them into the appropriate folders is tedious at best. The Web would be a good place to store and share pictures, but even the best photo Web sites have serious limitations in capacity, convenience, speed and security.

What we need here is a little more structure-an effective strategy for dealing with data that exist in many places at once. Fortunately, another part of our industry has already solved this problem. The solution is caching, a system for storing temporary copies of data along with status information. For many years, microprocessor architects have used caching technology to increase the effective performance of main memory in PCs and servers.

Most personal electronic devices should be treated as data caches, rather than independent storage subsystems. In this approach, one system is selected to store the master copy of each type of data. For example, the office server would become the home for all contact information. Laptops and PDAs would cache all or part of the contact database, while cell phones could borrow just the names and phone numbers of the most important contacts. Records updated remotely would be copied back to the database as soon as possible.

Using the caching approach instead of the local-storage model forces us to think about where our data really belong-but once we make this decision, data transfer can be automated. Each data item will know what it is, how important it is, and where it needs to be. We won't have to worry about whether our data are in the right place, or have to move them around manually. Every time a mobile device communicates with its primary host, data will move to where they're needed. When this communication is wireless and continuous, all of our data are always where we need them-and that's something we all could use.

Peter N. Glaskowsky is a senior editor of the Microprocessor Report. For more information on topics covered in this column, visit www.chipinsider.com. Send Peter e-mail at png@mdr.cahners.com.