Summer School 2014


The MSR India Summer School series, held in collaboration with the Indian Institute of Science, consists of lectures in a chosen area by leading experts from around the world. The aim is to introduce students and researchers to important new areas and the latest results and to provide a forum for Indian and international researchers to interact. The 2014 Summer School was held between June 23- July 02 at the Indian Institute of Science, Bangalore in the area of Wireless Networking.

In an increasingly mobile computing centric world, wireless communication provides the most basic form of connectivity to the internet. Over the past decade, wireless capacity has seen an almost Moore’s law like increase driven by practical realization of innovations such as MIMO, OFDM and LDPC Codes.

The 2014 Summer School focused on the latest advances in wireless networking such as full-duplex communication, rate-less codes, and white space networks, covering both theoretical and practical aspects of these exciting research areas. Participants also gained hands-on experience through targeted labs where they experimented with implementation using Software Defined Radio platforms such as WARP.

The School addressed both theoretical as well as practical aspects of the chosen area and is targeted at research scholars, faculty members, masters and senior undergraduate students.

The lectures were designed to offer self-contained introductions to chosen topics, leading up to some open problems for research. The speakers at the School were well known researchers in the field delivering lectures/ conducting workshops over the duration of the School.


Sachin Katti , Stanford University

Talk #1: Self-Interference Cancellation: Theory, Design and Implementation

Talk #2: Self-Interference Cancellation: Full Duplex & Other Applications

Self-interference cancellation invalidates a long-held fundamental assumption in wireless network design that radios can only operate in half duplex mode on the same channel. Beyond enabling true in-band full duplex radios that can transmit and receive on the same frequency, which effectively doubles spectral efficiency, self-interference cancellation tremendously simplifies spectrum management. It will unify TDD and FDD systems into one common standard, and it enables future networks to leverage fragmented spectrum, a pressing global issue that will continue to worsen in 5G networks. Self-interference cancellation offers the potential to complement and sustain the evolution of 5G technologies toward denser heterogeneous networks and can be utilized in wireless communication systems in multiple ways, including increased link capacity, spectrum virtualization, any-division duplexing (ADD), novel relay solutions, and enhanced interference coordination.

Bhaskaran Raman, IIT Bombay

WiFi: Speed Enhancements and Manageability

WiFi as a wireless technology for data is near-ubiquitous, and has even become part of laymen’s vocabulary due to its success. This talk will present the key technical enhancements over the different generations of WiFi over the last decade. We will look at research work on two specific aspects: data rate adaptation and power management in the context of the currently popular 802.11n. The second half of the talk will focus on an aspect which has thus far received relatively less attention: manageability of WiFi networks. I will present past work on manageability, dimensions where they fall short, and my own attempts to fill this gap.

Kyle Jamieson, University College London

Talk #1: Leveraging Angle-of-Arrival Information for Fine-Grained Indoor Location and WiFi Security

Phased array signal processing has long been employed outdoors in radar, underwater in sonar, and underground in seismic monitoring. But it has only recently made inroads indoors in the context of WiFi networks, where it must cope with strong multipath reflections, packetized data transmissions, and commodity hardware. I will begin by describing two systems my students and I have worked on: ArrayTrack (published at USENIX NSDI 2013), one of the first fine-grained indoor location systems, and SecureArray (published at ACM MobiCom 2013), one of the most effective physical-layer based security mechanisms for WiFi networks.

Talk #2: Partial Packet Recovery for Wireless Networks: From Software Defined Radios to Maranello

When link fading and noise overcome the error correcting codes of the wireless physical layer, wireless links often deliver packets with errors. These errors may be few and highly-localized, or they may be widespread across the packet, with just small pieces of the packet salvageable. Partial Packet Recovery protocols are specifically designed to correct packets rapidly in such situations. We’ll start by looking at PPR (ACM SIGCOMM 2007), one of the first partial packet recovery protocols, and continue with Maranello (USENIX NSDI 2010), a practical partial packet recovery protocol for the Broadcom WiFi chipset.

Talk #3: Rateless Codes: Spinal Codes

Spinal codes (ACM SIGCOMM 2012) are a new class of rateless codes that enable wireless networks to cope with time-varying channel conditions in a natural way, without requiring any explicit bit rate selection. The key idea in the code is the sequential application of a pseudo-random hash function to the message bits to produce a sequence of coded symbols for transmission.

This encoding ensures that two input messages that differ in even one bit lead to very different coded sequences after the point at which they differ, providing good resilience to noise and bit errors. To decode spinal codes, this paper develops an ap- proximate maximum-likelihood decoder, called the bubble decoder, which runs in time polynomial in the message size and achieves the Shannon capacity over both additive white Gaussian noise (AWGN) and binary symmetric channel (BSC) models. Experimental results obtained from a software implementation of a linear-time decoder show that spinal codes achieve higher throughput than fixed-rate LDPC codes, rateless Raptor codes, and the layered rateless coding approach of Strider, across a range of channel conditions and message sizes.

Kun Tan , Microsoft Research

Talk#1: Seven years of programming radio

Software radio is one of the most existing technologies in last decade. Being fully flexible and programmable, software radio has been widely used in research, education, instruments and measurement. In this lecture, I will share our experience to build a fast, real-time software radio system, namely the Sora platform. I will introduce the architecture and design philosophy of Sora and cover key optimization techniques that are crucial to fast signal processing in software. I will discuss how a better abstraction can greatly simply this optimization. Finally, I will introduce several directions that have been actively explored for the future software radio systems.

Talk #2: MIMO and dynamic spectrum access – a perspective from a system builder

Multiple Input and Multiple Output (MIMO) and dynamic spectrum access (DSA) are two latest technologies that significantly improve the wireless spectrum efficiency. MIMO allows multiple streams to be transmitted over the same spectrum, and therefore holds the promise to scale the wireless capacity linearly with the number of antennas, while DSA enables a more flexible spectrum allocation among heteronomous wireless technologies. I will share our experience in building a large scale multi-user MIMO system. I will introduce the challenges and solutions to scale a MIMO system. I will also discuss several open questions that may require more research from both theorists and system builders. In the second part, I will propose a spectrum virtualization layer (SVL) to support dynamic spectrum access as a first primitive for various wireless physical layer. I will discuss the architecture, the design and the implementation of SVL and how it can enable DSA with minor efforts.

Victor Bahl , Microsoft Research

Talk #1: White Space Networking – History, Issues and Update

On Sept. 23, 2010 the US Federal Communications Commission issued a landmark ruling which opened up UHF frequencies, formerly dedicated to television, for unlicensed use. Known as “white spaces” these frequencies enable the creation of a cellular network that can cover much larger areas than Wi-Fi can. These networks are supported by the IT industry, and research provides the muscle behind this push, transforming the concept into reality. To lead the way, we built the first operational, self-sustaining white-spaces network on our Redmond campus. In this talk I will discuss the feasibility of designing, deploying, and operating such networks. When white space networks become commonplace, companies with large campuses will be able to provide full coverage for their customers so that, as those customers move around a campus, they will continue to have connectivity, just as they currently enjoy within individual buildings. White space networking is considered by many as the next frontier of wireless Internet connectivity and the biggest opportunity in wireless communications since Wi-Fi and LTE.

Talk# 2: Cloudlets for Mobile Computing

Resource poverty is a fundamental constraint that severely limits the class of applications that can be run on mobile devices. This constraint is not just a temporary limitation of current technology, but is intrinsic to mobility. In this talk I will put forth a vision of mobile computing that breaks free of this fundamental constraint. In this vision, mobile users seamlessly utilize nearby computers to obtain the resource benefits of cloud computing without incurring WAN delays and jitter. Rather than relying on a distant “cloud,” a mobile user connects to a “cloudlet”, a micro datacenter on nearby infrastructure and uses it. Crisp interactive response for immersive applications that augment human cognition is then much easier to achieve because of the proximity of the cloudlet. While much remains to be done, the concepts and ideas introduced here open the door to a new world of mobile computing in which seamless cognitive assistance of users occurs in diverse ways at any time and place.

Ashutosh Sabharwal, Rice University

Antennas: Near and Far

In this short course, we will take a closer look at antennas and their diverse methods of use in wireless networks. We will first review the traditional viewpoint of using antennas 1) antennas which are near-by, like in MIMO systems, are good and 2) antennas which are far-away, like in multiuser systems, cause interference and are problematic. Then we will review the ideas which have emerged in last decade 1′) antennas which are near-by, like in full-duplex, are problematic, and 2′) antennas which are far-away, like in cooperatively coded systems, are good. The contrasting viewpoints will help us appreciate that our understanding of wireless systems is far from complete, and wireless continues to be an active and vibrant area of research.

Course Lab: While the lectures will focus on concepts and pointers to ongoing research, the course has a lab component. Using a custom developed WARPCloud, each student will be conducting brief experiments to appreciate the basic concepts in MIMO, Interference, Cooperation and Full-duplex.

Acknowledgement: The WARPCloud was developed and provided by Mango Communications, especially for the Microsoft Summer School attendees. Special thanks to Dr. Patrick Murphy, Dr. Christopher Hunter and Mr. Erik Welsh of Mango Communications, who took time off their busy work schedules and contributed towards the important mission of free educational tools for all.

Dhananjay Gore , Qualcomm

The 1000x Mobile Data Challenge

Mobile data traffic is exploding. Globally, the traffic has been doubling each year during the last few years and the industry is now preparing for an astounding 1000x increase. The solution to this formidable challenge is a combination of increasing the efficiency of existing assets, employing more resources in the form of small cells and spectrum, as well as adopting radically different ways of acquiring, deploying, operating and managing these resources.

The talk will focus on how extreme densification of small cells is the key to addressing 1000x challenge. We need to evolve small cells in all directions: all forms – micro, pico, femto, metro, relays etc.; all technologies—3G,4G, Wi-Fi, all integrated; deployments by operators as well as users. The densification begins with the existing spectrum and techniques that are available today, for example, optimizations such as “Range Expansion” are possible today with HSPA+ and in the future with LTE Advanced networks. Many more enhancements and new deployment models that are needed to reach the 1000x goal are being worked, on as part of standards and product solutions.

Venkatesh Ramaiyan, IIT Madras

Talk #1: Fixed point analysis of single cell IEEE 802.11e WLANs: Performance analysis

Talk #2: Fixed point analysis of single cell IEEE 802.11e WLANs: On decoupling approximation

We consider the vector fixed point equations arising out of a saturation throughput analysis of a single cell IEEE 802.11e (EDCA) WLAN. We study balanced and unbalanced solutions of the fixed point equations arising in homogeneous (i.e., one with the same backoff parameters) and nonhomogeneous networks. By a balanced fixed point, we mean one where all coordinates are equal. We are concerned, in particular, with 1) whether the fixed point is balanced within a class of users, and 2) whether the fixed point is unique. For IEEE 802.11 type WLANs, we provide a condition, based on the backoff parameters, for the vector fixed point solution to be balanced within a class, and also a condition for uniqueness of the solution. We then provide an extension of our general fixed point analysis to capture AIFS based differentiation and multiple virtual queues (supported in IEEE 802.11e EDCA); again a condition for uniqueness is established and simulations validate the analysis. The fixed point solutions are used to obtain insights into the throughput differentiation provided by different initial back-offs, persistence factors, AIFS and multiple virtual queues, for finite number of nodes, and for differentiation parameter values similar to those in the standard. Our simulations show that when multiple unbalanced fixed points exist then the time behavior of the system demonstrates severe short term unfairness (or multistability). Implications for the use of the fixed point formulation for performance analysis will be discussed.

Bozidar Radunovic, Microsoft Research

Wireless Programming for Hardware Dummies

Software defined radios are a powerful tool for experimenting with wireless PHY and MAC layers. At the same time, they are a challenging programming environment, given tight timing constraints imposed. A student who wants to venture in this area of research needs to master computer architecture and hardware, as well as numerous algorithms for signal processing and communication. In this lecture we will talk about Ziria, a programming language and a compiler that we have recently developed to simplify this task. Ziria is a high-level language, specialized for PHY design, that delegates most of the burdensome hardware optimization to the compiler and allows us to keep the code design clean and simple. We will walk through various building blocks of Wifi PHY design and show how to implement them in Ziria. At the end of the talk you should be able to understand the signal processing foundations of WiFi as well as to quickly implement and deploy your own PHY using Ziria. Ziria compiler is open sourced so you will be able to download it and play with the code yourselves. It currently supports Sora SDR platform but could be easily adapted to other similar platforms.

Sumit Roy , University of Washington

White Space Networking: Spectrum, Access & Co-Existence

The evolution of cognitive (secondary) networks to enable more efficient spectrum usage will rely on fast and accurate spectrum sensing/mapping, supported by a suitable architecture for data integration and model building. In the first part of the talk, fundamental aspects of the wide-area RF mapping problem as a grand challenge will be highlighted; and some recent work at UW that clarifies sub-system level trade-offs (between scan latency and channel status estimation accuracy, for example) will be described. Next, the evolution of a hybrid architecture – decentralized client-side sensing assisted database updating – is explored. Within this, model-based answers to fundamental questions such as “how much white space capacity is available” as a function of location for U.S. TV bands are developed. The talk will conclude with a description of current efforts for spectrum sharing (co-existence) just underway in the 3 GHz band (broadly) between different primaries (largely government operated communications such as military and non-military radars) and commercial networks (802.11 and 4G LTE).

A. Chockalingam , Indian Institute of Science

Algorithms for Large-Dimension Signal Processing in Wireless

Signal processing in large dimensions can bring several benefits in wireless communications. Detection of large-dimension signal vectors in large-MIMO channels is one example, where transmission and reception through large number of antennas can offer increased spectral efficiency and link reliability. Reception in channels with severe delay spreads (channels with a large number of resolvable multipaths like UWB channels) is another example, where rich multipath diversity gains are possible. This talk will focus on low-complexity algorithms suited for large-dimension signal processing. Channel hardening, a large-dimension effect, allows simple algorithms to perform very well in large dimensions.

The talk will specifically focus on algorithms based on Markov chain Monte Carlo (MCMC) techniques, message passing, and local search. The talk will also provide a status update on project NAVA, a Gigabit rate large-MIMO technology demonstrator project that IISc is working on jointly with DRDO and private industry.