Devices and Networking Summit 2015

Raffaello D’Andrea – ETH Zurich

We are at the cusp of a revolution: we can now create machines that adapt their behavior based on their environment and the results of their actions. The enablers for this revolution are sensing, communication, and computation technologies, innovative designs and novel mechanisms, and the feedback control algorithms that rule the machines.

These creations will have unprecedented effects on our lives – some welcome, others not. This talk will outline how we got here, where we are going, and the consequences.

Professor Sir Richard Friend- University of Cambridge

Pi-conjugated organic molecules and polymers now provide a set of well-performing semiconductors that support devices, particularly light-emitting diodes (LEDs) now widely commercialised in smart-phone displays and lighting, but extending also to field-effect transistors (FETs) and solar cells.

These are attractive materials to manufacture, particularly for large-area applications where they can be processed by direct printing, so that the cost of materials and processing can be very low. This practical success is made possible by breakthroughs in the understanding and engineering of the underlying semiconductor science.

Victor Bahl- Microsoft Research

Despite major advances in technology, resource poverty continues to limit the type of applications we are able to run on our mobile devices. The challenging issue is, this constraint is fundamental and not simply a temporary limitation of current technology. This talk will put forth a vision that breaks free of this constraint. In this vision, mobile users seamlessly use nearby micro datacenters to obtain the resource benefits of cloud computing without incurring wide area network delays and jitter.

Crisp interactive response for immersive applications that augment human cognition become easier to achieve. While much engineering and research remains, the concepts and ideas introduced open the door to a new world of disaggregated clouds in which seamless cognitive assistance for wireless users and new IoT applications can be delivered anyplace, anytime using the latest and greatest computer science technologies known to us.

Patrick Olivier – Newcastle University, Cecilia Mascolo – University of Cambridge, Nicolas Villar – Microsoft Research

Through the trends of the Internet of Things and Ubiquitous Computing, more and more devices are deployed around us that can sense what people do, and existing devices such as phones or wearables are augmented with new ambient sensing capabilities as well. Such devices are useful to both infrastructure owners and inhabitants alike, and the data they generate can be used for ‘smart buildings’ applications, for mass personalisation, for the users of shared spaces to communicate and collaborate, and so on. This session will explore this space through short presentations by researchers with experience deploying devices ‘in the wild’, and through a panel discussion.

Chair: Hitesh Ballani – Microsoft Research | Speakers: Mahadev Satyanarayanan – CMU, Flavio Junqueira – Microsoft Research, Arun Venkataramani – University of Massachusetts

This session will focus on the ‘cannot miss’ cutting-edge problem(s) in wireless systems design that researchers are working on and how their solution(s) will transform/impact the field. The aim is to create stronger ties between some of the top wireless and mobile systems researchers in academia and industry and to learn about the big ideas these researchers are working on.

Chair: Arjmand Samuel – Microsoft Research | Speakers: Jim Holbery – Microsoft, Jie Liu – Microsoft Research

Next generation electronic devices require unique sensing capabilities, low power, scalability, multiple form factors to integrate within multi-function network architectures. Material and process advances provides new electronic capabilities – this is especially true for electronics that are flexible, soft, textile, or printed incorporating brittle device materials, soft substrates, power harvesting and energy storage to aggregate data and transmit to the cloud. The physical form and computing interface of tomorrow’s IoT and mobile computing world is in a constant state of flux; this session addresses the fundamental challenges of future multi-disciplinary scientific and engineering research teams aimed at enabling next-generation devices.

Chair: Lucas Joppa – Microsoft Research | Speakers: Martin Wikelski – Max Planck Institute of Ornithology, Jie Liu – Microsoft Research, Jonathan Baillie – Zoological Society of London

We are now firmly living in the middle of the ‘Information Age’, and devices are central to our lives. From laptops to tablets, smartphones to game consoles, novel hardware is constantly entering the market and consumer demand has never been higher. But perhaps the true potential for new devices is bigger than putting music in people’s pockets, or advertisements in front of their eyes. If it isn’t already, science is set to become the driver of our times – from feeding an increasing population on this planet to potentially colonising another.

Scientists are not only key to developing new devices, but represent a key constituency pushing back the boundaries of what devices can be used for. This session will highlight the novel uses of devices in science, and consider some of the important, weird and wonderful discoveries only known to the world as a result of innovative device design. We will examine the question of how the scientific need for devices today might drive the consumer devices of tomorrow.

Chair: Stefan Saroiu – Microsoft Research | Speakers: Adrian Perrig – ETH Zurich, Alec Wolman – Microsoft Research, Peter Druschel – MPI-SWS

As more and more devices are connecting to the Internet, the security and privacy of their users are becoming more important than ever. The mobile nature of many of these devices raises new attack vectors and new challenges in providing adequate security and privacy.

This session will examine three different aspects of this problem: how to leverage the physical world to increase the security of online communication, how to protect data on a lost smartphone in the face of relatively unsophisticated memory attacks, and how to enforce user policies to control image capture by nearby mobile devices.

Chair: Bill Buxton – Microsoft Research | Speakers: Michel Beaudouin-Lafon – Université Paris Sud, Caroline Hummels – Eindhoven University of Technology, Jarnail Chudge – Microsoft

The ‘Next Big Thing’ is not a thing. It is neither a device, application, nor service. Rather, it is a reconceptualisation of the social relationships amongst these technologies. It is a change which helps us avoid the otherwise inevitability of our users being buried under the cumulative complexity of the increasing number of devices, applications and services that we, as an industry, create. Rather than pursue ‘Next Big Thing,’ it is far more important that we focus on ‘The Next Big Hurdle’ that this progressive accumulation of complexity presents.

To overcome this hurdle we need to see through different eyes, asking such questions as: What if things just worked? What if they just worked together? What if they worked together seamlessly? What if in working together, there was a significant increase in their cumulative value? What if in working together, there was a significant decrease in their cumulative complexity?

The founding premise of this session is that it would be far more disruptive, and beneficial to all, if the challenge emerging from each of these questions was achieved through a change in the relationship amongst existing devices, applications and services, rather than any ‘new thing’. The society of technologies will not evolve simply because their technology matures, but rather because their mutual relations are transformed. This panel will discuss the nature of just such a transformation.

Chair: Bozidar Radunovic – Microsoft Research | Speakers: Krishna Chintalapudi – Microsoft Research, Dina Katabi – Massachusetts Institute of Technology, Kyle Jamieson – University College London

Proliferation of mobile devices and more advanced applications have pushed the demand for wireless spectrum to the limits. Managing today’s networks and catering for the traffic demand becomes ever more difficult and requires new concepts in wireless network design.

At the same time, the new insights in wireless radio designs enable a string of entirely new applications, such as accurate device positioning and new ways of interacting with users. In this session we will give an overview of some of the recent advances in the area and illustrate how smart wireless can help smart devices.

Chair: Richard Banks – Microsoft Research | Speakers: Rob Girling – Artefact, Nikki Barton – Smart Devices UX Design, Nokia/Microsoft, Jon Rogers – University of Dundee

How do you design devices that people really love? As devices have started filling our homes and adorning our bodies the expectations of our customers have risen. It is no longer enough that our devices are technologically interesting. Neither is it enough that they are simply usable. The devices in our lives now need to be carefully crafted to be compelling, responsive and an extension, through the scenarios that they enable, of the way we see ourselves. This takes careful design. In this panel session we’ll hear from three luminaries of the design community who span the worlds of software interfaces, the Internet of Things and physical computing. They’ll give examples and insights into what it takes to design and build compelling device experiences, before opening the session up for discussion.

Chair: Christos Gkantsidis – Microsoft Research | Speakers: Dina Papagiannaki – Telefonica, Renata Teixeira – LIP6, Thomas Karagiannis – Microsoft Research

Home networking is rapidly transforming to accommodate new experiences backed by cloud services. Alongside video streaming that has been hogging wireless and access links for a long time, new services that, for example, simplify music consumption, photo sharing, document editing, and backing-up put even more strain on the home network’s resources.

This session brings together experts that will shed light on the configuration, problems and challenges faced by home networks. It will also examine the consumption of cloud services in home environments, and outline opportunities for improving the home networking experience.

Desney Tan and Ali Alvi – Microsoft Research and Shwetak Patel – University of Washington

There is an emerging trend to instrument the world with sensors that provide insight and allow us to improve the health and function of various systems. For example, cars now use many sensors for on-board diagnostics shifting us from reactive crisis-driven repairs to proactive preventive maintenance. Similar trends pervade our homes, workplaces, and outdoor environments. We must apply this approach to the system we arguably care the most about – our bodies. We will open the session with a high level overview of wearable health technology. We will describe our early research on exercise and activity sensing, which has now been incorporated into the Microsoft Band. We will then demonstrate capabilities of the Band, as well as present example applications built with the Software Development Kit. We will close with some of our current work, pushing beyond consumer health and fitness into clinical health sensors.

Bozidar Radunovic and Dimitrios Vytiniotis – Microsoft Research

Software-defined radio (SDR) brings the flexibility of software to wireless protocol design, promising an ideal platform for innovation in wireless. However, most of the currently available code for existing SDR platforms is too slow to meet the requirements of modern wireless standards. Implementing these requirements demands careful hand-tuning of low-level code, and the code becomes difficult to understand, modify and reuse, undermining the advantages of software. In this tutorial we will show you Ziria, a new platform for rapid prototyping of wireless physical layer (PHY) on SDR. Ziria allows a programmer to express a wireless PHY program (protocol descriptions and signal processing algorithms) at a high level, while the compiler takes care of low-level optimisations under the hood. Ziria’s source code is easy to understand, reuse and modify, and the executable is fast. Each participant will get a Blade-RF SDR platform and through hands-on exercises, will build a WiFi sniffer. All of the code used in the tutorial is open source (including Ziria itself) and the BladeRF hardware is available for you to take away.

Peli de Halleux – Microsoft Research and Dean Mohamedally – UCL

Hardware 2.0 is upon us: cheap micro-controller boards like Arduino have gained massive adoption in recent years. Paired with 3D printers, cheap sensors and actuators, it allows anyone to prototype the next hot gadget. And yet, the maker will have to learn a soup of software language and framework to build a connected input/output controller solution: C++ for the micro-controller code, HTML + javascript for the client, some backend language and a communication layer to interact with the devices. In this tutorial, we will show an attempt at unifying the compilation of web server code, rich client and embedded firmware under a simple mobile friendly language and IDE. Attendees will be given an Engduino, a Micro SD card loaded with TouchDevelop and a Raspberry Pi to take away.

Arjmand Samuel and Ratul Mahajan – Microsoft Research

Today we live in a world where computing is moving beyond tablets and smartphones, to everyday objects comprising of connected devices and sensors. We call this phenomenon the Internet of Things (IoT). While this new world of the Internet of Things is promising new opportunities for innovation and creativity, it is also exposing deep research challenges which need to be addressed. During this session, key research challenges associated with large scale deployment of the Internet of Things will be highlighted, leading to an introduction to the Microsoft Research Lab of Things, which allows deployment of IoT prototypes and research studies at scale. Attendees will be able to work with sensors and devices, including a multi-sensor and smart switch, interconnected over a ZWave network to enable interesting IoT scenarios. These devices will be available to take away from the Summit.

David Sweeney, Steve Hodges and Nicholas Chen, Microsoft Research

Taking inspiration from old chemical photographic processes, we are creating novel displays which are both analogue and digital in nature. Certain display driving architectures have become virtually universal, however their construction limits the potential usage for designers. Here we aspire to treat displays as a material rather than unyielding and fragile devices. We can now imagine displays which are resilient, of low power, complexity and cost and apply them to previously unimagined forms and situations.

Patrick Olivier, Vasilis Vlachokyriakos and Karim Ladha, Newcastle University

PosterVote is an artifact that allows sustainable electronic voting by dropping the development and maintenance costs, while increasing the potential for social movements to engage in action and for communities to support and respond to such action. PosterVote can be used by citizens to collect data, in the form of opinions, to support local communities. It consists of two parts; the printed poster with the question and answers and the hardware which is a thin, flexible circuit board with 5 buttons and 5 LEDs to be placed at the back of the poster.

Patrick Baudisch and Stefanie Mueller, Hasso-Plattner Institute

Computer science and mechanical engineering are about to unite. Through the use of 3D scanners and 3D printers, we will solve mechanical problems with the tools and with the effectiveness of computer science. The role of our group is to drive this unification process, in particular by creating and re-purposing fabrication machines and haptic machinery. In this demo, we present one facet, namely how to make 3D printing fast, so as to allow for interactive use.

Ali Alvi and Tony Andrews, Microsoft

In the demo, we will demonstrate how to use Microsoft Band SDK to write applications that communicate with the Band. We will show how to set up your developer environment and connect to Microsoft Band in code. Also we will show how to subscribe to various sensors on the Band and talk about what data those sensors report and how it can be used in various algorithms and inferences.

Peli de Halleux, Microsoft Research and Dean Mohamedally, University College London, Atia Rafiq, and students from University College London

More and more devices are entering our lives: devices we can use, such as phones and tablets, devices we can program such as Arduinos and other IoT devices. TouchDevelop, which enables the creation of apps from any device, was extended to allow the creation of device firmware. Leveraging the beginner friendly TouchDevelop environment, one can use devices to program devices. A team from UCL will share their experience in building and using the Engduino board for teaching.

Day 1: Paul Thomas Microsoft UK and Eliezer Vogel University College London; Day 2: Arjmand Samuel and Ratul Mahajan, Microsoft Research

We present an innovative and interactive architecture for IoT platforms comprising of a middle aggregation layer constituted of a Captain Device running Windows Operating System and the Lab of Things (LoT) Home OS platform. The demo show cases a generic Bluetooth extension of the LoT platform using the UCL Engduino and Android smartphone as example devices enabling cloud data storage and cloud based scenario management.

Bozidar Radunovic and Dimitrios Vytiniotis, Microsoft Research

In this demo we will show how to implement a WiFi receiver in Ziria, a new platform for rapid prototyping of wireless physical layer on Software Defined Radio, that is able to receive WiFi packets. We will connect a BladeRF board to a PC and display packets received from a nearby WiFi access point at line speed.

Christian Moller and Jim Holbery, Microsoft

Recently the Applied Science Group has developed touch sensor arrays constructed from unique textile architectures that demonstrate multi-touch sensor capabilities. This new architecture enables designers to embed touch functionality within fiber composite structures without compromising the structural integrity of the hardware. We will show designs that have been proven to deliver multi-touch performance from a textile platform, and will demonstrate these functional touch sensors embedded within structures providing unique functionality to next-generation electronic devices.

Jim Holbery and Christian Moller, Microsoft

Using drop-on-demand inkjet printing, we have produced the first reported touch sensor from transparent organic conducting polymers. The Microsoft Applied Science Group has designed, printed, and tested transparent multi-touch sensor arrays from unique material forms that include organic conducting PEDOT:PSS polymers and inks comprised of multi-wall carbon nanotubes. The sensitivity of the sensor arrays have proven to be comparable to conventional materials but offer low haze combined with the possibility of deployment on flexible substrates.

Jarnail Chudge, Microsoft UK

Cities unlocked is a project that assists people with visual impairment to navigate through urban areas. It uses a bone conducting headset and a mobile phone app to generate audio signals based on the user’s head position, to guide them towards a target. This is a highly collaborative effort between Microsoft and external partners, including UCL. Microsoft Research contributed to the algorithms for the soundscape technology, made the hardware modifications to the headset and designed the user study.

David Chu, Microsoft Research

Head-mounted Displays (HMDs) are a promising way to achieve virtual reality (VR). For truly immersive VR, three properties are essential: quality, responsiveness and mobility. Existing systems fail to achieve all three. We demonstrate Irides, a stereo HMD system that aims to achieve all three. Irides offloads rendering work to powerful remote GPUs for high quality visuals, adapting previous work on speculative execution to overcome any motion-to-photon (including network) latencies.

Steve Hodges, Microsoft Research and Yoshihiro Kawahara, University of Tokyo

Circuit stickers let you build electronic prototypes by sticking components like LEDs, sensors and logic onto a substrate which wires them together. We use paper and conductive silver nanoparticle ink to make this wiring layer – it’s cheap and easy to create using a felt pen or a domestic inkjet printer. The technology has potential for design exploration, research prototyping, education and for hobbyist projects.

Nicolas Villar, Microsoft Research

Picco is a tiny networked device that displays simple animations, which are created on a dedicated tablet app. Picco was designed to support playful messaging that would place minimal demands on users: the device’s small size makes it unobtrusive, and messages are delivered through a queuing system that discourages the sending of time-critical content. Alerts are subtle and can be acknowledged with a simple tap.

Sergey Legtchenko, Microsoft Research

Datacenter hardware is changing. Thanks to technologies like Systems-on-a-chip and resource disaggregation, the number of servers per rack is likely to increase from 40 today to about 1000 in the near future. We demonstrate XFabric, a rack-scale network designed to support this increasing compute density. XFabric dynamically adapts to the workload traffic by reconfiguring its network topology at the physical layer. XFabric performs packet switching over physical circuit switching and uses low cost commodity ASICs (application-specific integrated circuits).