January 19, 2021 | Written by: Katia Moskvitch
Categorized: AI | Awards and Prizes | Systems
Share this post:
A chat with one of IBM’s leading computer scientists and 2020 ACM Fellow Chandra Narayanaswami
We now take smartwatches for granted. But did you know that very early models go back two decades?
Between 2000 and 2001, IBM Research made headlines when it unveiled an internet-enabled designer watch that ran on the open-source operating system Linux. Dubbed WatchPad, its aim was to demonstrate the capabilities of the then-novel OS for mobile and embedded devices — complementing other efforts to adapt Linux for devices ranging from workstations to large enterprise servers. The futuristic watch featured Bluetooth communication and a VGA (640-by-480 pixel) OLED screen, ran Linux version 2.4, and even enabled mobile payments.
Chandra Narayanaswami, Distinguished Research Staff Member, Member IBM AoT, Member IBM Industry Academy, IBM Research
That’s not the only early mobile system that came out of IBM in the early 2000s. Four years after the WatchPad, the same team of researchers led by computer scientist Chandra Narayanaswami created a portable virtual computer — the SoulPad. It was the time before smartphones or slim and light laptops, and cloud computing was just nascent. (The first cloud platforms became publicly available only in 2006.) Meanwhile, IBM’s pocket-size device allowed users to carry their desktop around with them and re-create their home or office computer on any PC via a USB or FireWire port.
Fast-forward to today — the two ‘Pads’ main designer, Narayanaswami, has just received one of the highest recognitions in computer science. He’s been elected a 2020 ACM Fellow “for design and development of the Linux Watch and SoulPad, which influenced wearable and mobile systems.”
Chandra, please tell us more about these two projects that influenced — and to a degree kick-started — the entire wearable and portable device revolution?
It was the late 1990s, and along with the advent of the internet, there was a proliferation of non-programmable fixed purpose gadgets, I am sure many will recall pagers. At the same time, several ambitious projects seeking to connect these various devices had just gotten underway in academia. Back then, IBM had industry-leading notebook computers, and a few of us at IBM Research wanted to rethink computing devices for the emerging landscape of pervasive computing.
We chose to build a watch computer specifically because of the challenges imposed by the small form factor. The WatchPad, as it was called, incorporated many technological firsts and significantly pushed the boundaries in hardware packaging, power management, and embedded software to become the tiniest wireless device to run the standard Linux operating system, wireless software stacks, and graphics libraries.
IBM Watchpad, the internet-enabled designer watch running Linux from 2000
On top of that, the WatchPad demonstrated many unforeseen capabilities in a wearable and mobile device. It was the first wearable with biometric authentication and password vaults, Bluetooth wireless-based control and collaboration functions, and mobile web services, such as payments. A wireless body-worn hub, derived from the watch hardware in 2003, was in fact one of the earliest devices to sport mobile health monitoring. The Linux WatchPad won the Red Dot Design and Best of PC Expo awards and appeared on the cover of IEEE Computer.
As for SoulPad, this project fulfilled a long-cherished vision of the mobile and ubiquitous computing community to walk up to any computer, personalize it, and use it as one’s own.
In essence, SoulPad allowed users to start a computing session on one PC, stop it, and resume it later on another PC with just a mobile device like a pocket drive or an iPod — the “Soul.” It did this by combining an auto-configuring operating system and a virtual machine that were stored on the mobile device. We also laid out several of the improvements needed in virtual machine technology, application and system software design, and software licensing for this style of suspend/resume operation. Our paper describing this work won the Best Paper award in ACM MobiSys 2005. Commercial solutions such as U3, MojoPac, Windows To Go, Ceedo, and so on appeared soon after.
How did you get into this field of research?
IBM made amazing ThinkPads in the late 1990s, and the company looked to IBM Research to lead the way in building the future of personal computing. During this time, I was exposed to researchers working on multiple systems areas, ranging from efficient power management and thermal cooling to novel input devices.
But it was becoming clear that even the amazing ThinkPads were too big and heavy. Around this time, I started using early portable GPS units and digital cameras, and attended some of the early IEEE Symposiums on Wearable Computing. These experiences convinced me that we could unlock a lot of potential by making devices more portable and connected. I was hooked on mobile computing. Fortunately, the field was just emerging and there was a lot a computer scientist could do.
Fast forward 20 years — I have since changed my area of research and now I lead the Supply Chain Theme for IBM Research, working with AI. To my pleasant surprise, the work I did on the WatchPad continues to be relevant here, for example in the case of CryptoAnchors. This project incorporates tiny devices with a secure computing platform and sensors able to capture data from the physical world and report it to a blockchain ecosystem to document and protect our supply chains.
What especially excites you today about the field of mobile and wearable computing?
The COVID-19 pandemic has actually given rise to several interesting mobile applications, such as contact tracing and worker safety monitoring. These are great examples of wide-scale application having societal impact across the world. New uses continue to emerge for this field. Devices are becoming more capable and novel sensors are emerging leading to more possibilities at home and work.
I personally think there is a lot more to come from synergies between pervasive devices and the cloud. For example, I imagine a future where I could dictate a task to my home automation system to fly a 5G-connected drone to inspect my roof or my backyard. And then, it would produce an itemized report of issues that need to be fixed, along with suggestions for contractors to perform those repairs.
Talking about your current field of work, AI in Supply Chains. What are the most cutting-edge developments?
We are seeing more and more the need for intelligent self-correcting supply chains because of the scale and speed at which they need to operate. For example, what specific sequence of actions should a Chief Supply Chain Officer take in response to an impending hurricane, a potential supplier disruption, or to increasing incidence of customer dissatisfaction? Recent advances in AI — including deep learning and reinforcement learning, explainable AI, availability of real-time data from business processes and IoT devices, improved connectivity, and cloud-based computation — will help us get there.
Several challenges also arise in quickly finding and reconciling the relevant data needed for a particular problem. I call this the “data supply chain,” where the production, sourcing, quality, pricing, delivery, ratings, disposal, and recycling of data is treated with the rigor of a traditional supply chain. Operating efficiently at scale and speed continue to remain a challenge.
To help make progress, our research teams are using advanced optimization and AI techniques for faster and more accurate demand forecasting, inventory positioning and replenishment, and resource optimization. We are developing privacy-preserving platforms for sharing and analyzing data in multi-enterprise supply chain networks, applying workflow automation to improve efficiency and speed, building carbon calculators for reporting and improving supply chain sustainability, and so much more.
What’s the next step in your research and how can it benefit society once you achieve those future results?
People are starting to ask if quantum computing is ready for the optimization problems encountered in supply chains and whether it can help solve some of the above challenges. We need to find some of these answers.
We hope that our work will pave the way for a world with more intelligent and sustainable supply chains where people spend more of their time on interesting tasks and leave the repetitive, “boring” work to machines.
Inventing What’s Next.
Stay up to date with the latest announcements, research, and events from IBM Research through our newsletter.