Looking Back: 5 Technologies That Have Paved The Way For Mobile Computing

We've come a long way since 1984. That was the year when IBM debuted its Portable PC. Well, at 30 pounds, it was about as portable as an old-fashioned sewing machine. And there was the issue of batteries: The machine didn't have any. Instead, it relied on 110-volt AC power at all times. Its processor, a then state-of-the-art Intel 8088, ran at just 4.77 MHz, though, to be fair, that was perfectly matched to the system's pair of 5.25-inch, 360-KB floppy drives and 9-inch amber CRT. For all this--plus serial and parallel ports--mobile-computing pioneers had to plunk down some major cash; to be exact, $4,225.

What a difference 20 years makes. Today's business notebooks typically weigh no more than 5 pounds. Their processors are several million times more powerful. Storage options include an 80-GB internal hard drive paired with a CD-RW or DVD drive. Thin, 14-inch LCD displays can display 24-bit true color movies and 3D models. Rechargeable batteries last up to five hours, even longer with auxiliary battery packs. Communications are relatively easy with both local and remote networks at up to 54 Mbps. In short, mobile users can do just about any kind of computing they could do on their desktops--and all for less than half the price of IBM's 1984 portable.

"A value threshold was reached in the early 1990s," recalls Howard Locker, chief architect for desktop and mobile development at IBM's Personal Computing Division. "When mobile machines moved from luggable to useful, becoming powerful enough to do real work, it was a magical time."

Indeed, during the 1990s, processor power doubled every 18 months. LCD displays grew to comfortable reading sizes and resolutions. Hard drives shrank to 2.5 inches. Storage capacities leapt from megabytes to gigabytes. Batteries doubled in energy storage capacity. And buyers responded enthusiastically.

From 1995 to 2004, notebooks' share of the total business PC market grew from 18 percent to 29 percent, according to market analyst IDC. And last year, more than 31 million business notebooks were shipped, a 60 percent increase over 2000, driven in large part by major breakthroughs in wireless communications.

How did we get here? In large part, courtesy of five core technologies that have both enabled the modern mobile PC and been enabled themselves by rapidly growing demand for mobile systems: processors, displays, hard drives, communications and batteries. Their story is also the story of mobile computing.

Processor Power

Ever since mobile-computer makers cut the cord, they've struggled to match desktop performance while also minimizing their machines' size, energy consumption and heat output. Back in 1983, the Epson HX-20 laptop got 50 hours per battery charge, but its Hitachi 6301 CPU ran at only 0.6 MHz. Since then, advances in microprocessor design have helped to close the gap. Intel and AMD have been locked in a virtual arms race that has benefited users above all. One of the duo's first innovations involved consolidating functions onto core processors. In 1989, Intel added floating-point instructions to the 80486 as a way to replace math co-processors, thus lowering both power needs and weight. Nearly 10 years later, in 1998, Intel was the first to offer on-die Level-2 cache. Next, AMD moved its memory controller onto the processor die. Both these designs boosted processor performance without requiring additional power. They also enabled the smaller chip packages needed for ultra-slim notebooks. Later came faster processing of graphics and multimedia data without separate accelerator circuitry, a trick enabled by both Intel's MMX and SSE2 instruction sets, and AMD's 3DNow! technology. These chipsets integrated audio, video and other I/O controllers.

More savings were realized by fine-tuning power levels to meet demand. In 1989, Intel's SL technology had given the BIOS coarse control over power levels based on inactivity timers. Unfortunately, it would also often turn off a device just before a program needed it. That was remedied by 1996's ACPI spec, which enabled intelligent power management based on the operating system's awareness of running applications. That, in turn, led to improved CPU power management.

More recently, Intel's SpeedStep technology, announced in 2000, toggles mobile processors between full- and low-power modes when a notebook switches between wall-outlet and battery power. AMD, in turn, responded with its PowerNow! technology, which adjusts CPU voltage and speed in response to the needs of software applications, using five voltage levels to Intel's two. Then, in 2003, Intel announced Enhanced SpeedStep, which added similar dynamic speed switching and ultra-low power levels to the Pentium M.

In fact, Intel's Pentium M is noted for being the first processor designed for mobile computing from scratch, rather than being derived from a desktop processor. The M's overall logic design has been optimized for mobile power and thermal constraints. Its high instructions-per-cycle rate lets the processor do more work in fewer energy-burning clock cycles. And the processor's L-2 cache and processor bus are divided into parts that can be powered off individually when not needed. Micro-op fusion reduces traffic in the CPU pipeline by 10 percent, saving yet more power without a loss in performance.

AMD hasn't been sitting idly. The company implemented the antivirus Data Execution Prevention feature of Windows XP Service Pack 2 a full year before Intel. It also led the way to introducing 64-bit chips for notebooks, announcing low-power versions of its Mobile Athlon 64 last year, and this year releasing Turion, a processor for the ultrathin notebook market.

Looking ahead, the twin issues of power consumption and heat dissipation are likely to become the overriding design criteria for processor architects. That's because chip fabricators will continue to cram more transistors into ever smaller spaces.

What's On the Tube?

The bulky, power-hungry CRT screens of early mobile computers couldn't last. Liquid-crystal display (LCD) technology has quickly become the norm. LCD display technology's early days--for example, the blocky black pixels of Radio Shack's TRS-80 Model 100--have given way to 17-inch DVD-ready screens. And the LCD itself has gone through changes.

That old TRS-80 Model 100's LCD employed a technology known as passive-matrix. The screen incorporated a grid of wires--electrical current-activated pixels at the wires' intersections using low-contrast twisted nematic (TN) crystals. Contrast was soon improved with the innovation of super-twisted nematic (STN) crystals. Still more improvement came with film-compensated STN (FSTN) displays, which enabled black-and-white displays and wider viewing angles. FSTN displays also had the advantage of being inexpensive to build, making them attractive for low-cost PDAs and mobile phones.

But LCD displays really took off with the advent of active-matrix technology. That innovation brought the large, sharp, full-color and full-motion screens we know today, though at a hefty price. But customers quickly decided the advantages of big color displays outweighed their cost, and the rest, as they say, is history.

In fact, the notebook market really took off in 1992, points out Howard Locker, chief architect for desktop and mobile development in IBM's Personal Computing Division. That year, IBM introduced the industry's first 10.4-inch color display on a mobile computer, the IBM ThinkPad 703c. Today, notebook displays are as large as 17 inches, although 14 inches remains the most popular size among business users.

Active-matrix technology brings other benefits, too. It gives designers precise control over individual pixels. Screen response time--the lag between switching on voltage and its visible effect on a pixel--can be as little as 12 milliseconds, compared with 150 ms for other screen types. And active-matrix screens can handle a higher current without crossover leakage, achieving greater contrast and brightness with little ghosting. However, active-matrix LCDs have only one problem--and for mobile users, it's a doozy: They are a battery's worst enemy.

On a typical notebook, the 14-inch TFT display alone consumes fully one-third of the system's total power. That's more than is consumed by the CPU, chipset and hard drive combined. Yet manufacturers have kept the power requirements of LCDs fairly constant over the past decade, even as the average TFT display has grown in size from 10 inches to 14. That impressive feat has been accomplished primarily with a variety of power-saving innovations.

To be fair, conserving display power is a cooperative job. Screen fabricators now use a polysilicon substrate that lets a screen work at lower voltage levels with smaller transistors. This also allows more light to pass through the screen, reducing the backlight's power requirements, which can account for half of all the power the screen needs. Similarly, processor makers have created chipsets that can check the brightness and contrast of images, then adjust backlighting accordingly, a neat trick that can reduce backlight power by as much as 25 percent. And notebook makers now offer ambient light sensors that measure the light levels around the computer, then brighten or dim the screen's backlight accordingly.

What's next for mobile displays? One expected breakthrough is screens that use a technology known as Organic Light-Emitting Diodes (OLED). Like old-fashioned CRTs, newfangled OLEDs are self-illuminating. But because they need no backlight, they can be thinner, lighter and less power-hungry than today's displays. They can also switch on and off almost instantly, and can be viewed from almost any angle. Even better, current prototypes have gotten as large as 19 inches diagonally. Today, building OLED screens is still too costly, say industry insiders, who expect the cost problems to be solved in five to seven years.

Hard Drives, Hard Bargains

Early mobile computers limped along on floppy drives. The Osborne 1, for example, sported twin floppies that each held just 91 KB. IBM's early Portable PC upped that to a pair of 360-KB floppies. But hard drives with 10 MB or 20 MB of capacity didn't appear in portable computers until about 1986.

Those early hard drives were obese gluttons, recalls Bill Healy, senior vice president of product strategy and marketing at Hitachi Global Storage Technologies.

"Early [mobile] hard drives had 5.25-inch platters and were about the size of two bricks," Healy says. "Drives were second only to displays as power hogs."

In fact, a 3.5-inch 60-MB drive--state-of-the-art circa 1991--devoured 2 watts while idling, Healy says. In comparison, today's miniaturized hard drives sip only 0.65 watts at rest.

Since then, advances in both storage software and hardware have helped to conserve electricity. Hitachi, for one, added Adaptive Battery Life Extension (ABLE) microcode to its drives' processors to track drive activity and selectively power down various parts, optimizing power conservation and driving availability. ABLE parks the read/write head not only to save its power, but also to reduce aerodynamic drag that would burn more energy.

Notebook hard drives shrank to 2.5 inches in diameter during the early 1990s and remain there today. This form factor represents a good compromise between size and angular velocity near the platter's rim on the other; the latter impacts seek-time and data-access speeds. Yet Healy expects notebook makers to gradually adopt a new generation of 1.8-inch and even 1-inch drives, mainly because the smaller drives are both lighter and more resistant to bumps and drops.

Innovations in drive components have also improved performance. For example, two years ago, Hitachi shrank the slider in its Travelstar 7K60, a 2.5-inch drive, by some 30 percent. (The slider holds the read/write head and moves across the disk, like the arm of a phonograph turntable.) That innovation, in turn, improved the drive's data-transfer rate by 15 percent and reduced its average access time by 20 percent.

Drive capacity has grown, too--even faster than the storage needs of most business users, Healy says. Consumers, eager to store music and video files, have pushed notebook-drive capacities beyond the 100-GB mark. For business users, however, the sweet spot is in the 40- to 80-GB range.

The shock resistance of today's hard drives also has increased several hundredfold. Back in 1991, a force of 5 g's was all most drives could tolerate--the equivalent of a mere elbow nudge. Today's drives can tolerate up to 300 g's while operating, and 1,000 g's when the heads are parked. The latest innovation is an accelerometer, a device that detects when a notebook is falling. When it senses a drop, the accelerometer sends instructions to the drive to park its heads--hopefully before the machine hits the floor.

Communications Commotion

The dream of mobile users is anywhere, anytime connectedness. While we're not there yet, we've come a long way, and mostly in the past five years.

In the Before Internet era, dial-up modems led the way. But modems were slow, communicating at only 9,600 bps in 1989 and maxing out at 56 Kbps in 1998. Nationwide networks of local-access phone numbers from services such as Prodigy and CompuServe helped reduce long-distance charges, but only in major metropolitan areas. Still, mobile modem users had to deal with multiple dial-up networking configurations, digital PBXs that blocked analog signals and being tied to a wire in a room.

Relief came in 1989 with the first commercial wireless data services. For example, Ardis and Mobitex offered packet-switched networks, while cellular phone companies provided circuit-switched services. Both were slower than dial-up modems of the time, typically just 4.8 Kbps to 9.6 Kbps. But, today, Verizon offers 2 Mbps data service in 32 cities, and other carriers are rushing to compete.

At about the same time wireless networks were starting, Ethernet came to mobile computers. The year 1988 saw the Pocket LAN Adapter from Xircom. This device plugged into a notebook's parallel port to enable 10-Mbps LAN connections via a cable. The product proved so popular that by 1992, most mobile computers featured integrated Ethernet ports in 10-Mbps, 100-Mbps and, eventually, 1,000-Mbps versions. But they still needed a wire.

The PC Card and the later, faster PCBus standards, spawned myriad slide-in cards incorporating a choice of Ethernet and a modem, modem-to-cellular phone connector, and even cellular modems and phones on a card. Ethernet was fast but wired. Wireless was mobile but slow. Finally, in 1999, they married.

The 802.11b wireless Ethernet standard launched the Golden Age of mobile communications. The technology offers speeds of up to 11 Mbps at 100 feet to 300 feet from a base station, enabling productivity equivalent to wired LAN access. Then the 802.11a standard raised the bar to 54 Mbps over shorter ranges. The 802.11 standard--better known as Wi-Fi, short for Wireless Fidelity--has continued to evolve.

At the same time, bulky antennas quickly yielded to PCBus cards. By 2003, nearly 90 percent of all notebooks were shipped with a Wi-Fi card. And Wi-Fi has spawned a new industry of wireless ISPs that provide "hotspots" of Wi-Fi access in airports, cafes, hotels and other public spaces.

The future looks even faster. An 802.11n standard is in the works to support more than 100 Mbps. Cellular operators are demonstrating 1 Gbps in lab environments. And Wi-Fi hotspot coverage continues to expand, in some cases covering entire metropolitan areas. Cellular service, however, is likely to grow more slowly, due to both the multibillion-dollar costs of upgrading cellular networks and wrangles over international standards.

The Battery Barrier

With lightweight notebooks our constant companions, battery life is seen as the killer function. Unfortunately, today's battery technology has neared the limits of physics, and longer battery life is probably unfeasible until the industry switches to a new battery technology. Some of those technologies include "smart" batteries, lithium-polymer (Li-poly) and fuel cells.

Tandy's pioneering TRS Model 100 used four AA alkaline cells. Apple's original Mac Portable held a lead-acid battery so heavy it would outweigh many of today's ultra-thin notebooks. Throughout the 1980s, rechargeable nickel-cadmium (NiCd) batteries were most common. They were followed, in turn, by less toxic nickel-metal hydride (NiMh) batteries, which stored 40 percent more watt-hours per kilogram than NiCd. But since 1993, the standard has been the lithium-ion (Li-ion) battery, which offers roughly triple the energy density of NiCd.

Battery makers have recently started to add electronics that create "smart" batteries. A smart battery essentially monitors itself. It notes its state-of-charge, remaining voltage, drawn current levels, internal temperature and other battery-health factors. These batteries can also communicate with the systems they power. For example, a smart notebook battery could alert the computer, say, 10 minutes before it needs a recharge.

On the next frontier of battery technology is Li-poly. This technology could replace Li-ion within the next few years. Li-poly's energy density is slightly lower than Li-ion's. It's also cheaper to manufacture. Further, Li-poly batteries are lighter, since they don't need a metal case. Their shapes are also less constrained, so they can be made to better fit notebook designs. And from a safety perspective, Li-poly batteries are less susceptible to overheating and electrolyte leakage.

Assuming Li-poly technology catches on, it may be the last battery your notebook ever sees, as fuel cells catch on. Unlike batteries--which store electricity--fuel cells combine hydrogen and oxygen to produce water and electricity. The energy density of a fuel cell is from five to 10 times that of a Li-ion equivalent, which means it can be dramatically smaller.

But until the battery's equivalent of a Fountain of Youth is discovered, these innovations point to the future's most likely direction.