New Directions in Networked Systems Design

Prof. Kyle Jamieson (opens in new tab), University College London

Abstract: Location systems are key to a rich experience for mobile users. When they roam outdoors, mobiles can usually count on a clear GPS signal for an accurate location, but indoors, GPS usually fades, and so up until recently, mobiles have had to rely mainly on rather coarse-grained signal strength readings for location. I will talk about the design and experimental evaluation of ArrayTrack, an indoor location system that uses Angle-of-Arrival information at access points to track wireless clients in real time as they roam about a building. Our results show that ArrayTrack can pinpoint 33 clients spread out over an indoor office environment to within a median 25 cm location accuracy.

Nathan Farrington (opens in new tab), University of California San Diego

Abstract: In this talk, you will learn why copper electrical wiring can never effectively be used to build high-speed data center networks at 40G and up, and no matter how much as computer scientists we kick and scream, we will be forced to learn to stop worrying and love optical communications.

Mohammad Alizadeh (opens in new tab), Stanford University

Abstract: The datacenter increasingly resembles a warehouse-scale computer: a large collection of computing and networking resources that work in concert to efficiently deliver superior performance. For this vision to fully materialize, transport in the datacenter (the plumbing) needs to vastly improve. I will discuss some of the opportunities and challenges in this space and recent research that aims to address them.

Emin Gun Sirer (opens in new tab), Cornell University

Abstract: You’ve heard of the NoSQL revolution and the CAP Theorem. In this talk, I will tell you why CAP is not what it’s cracked up to be and describe a revolutionary new architecture for NoSQL data stores that are consistent, available in the presence of partitions that affect up to a threshold of nodes, scalable, and above all, fast.

Prof.Vyas Sekar (opens in new tab), State University of New York Stony Brook

Abstract: Middleboxes (firewalls, IDSes, proxies, WAN optimizers, and the like) have long been an integral part of operational networks, but have traditionally been treated as second-class citizens in the research community. There is growing recognition of the need to bridge this disconnect in both camps. I will describe my recent research related to the design, implementation, and management of middle-boxes and discuss some challenges and opportunities in integrating middleboxes with SDN mechanisms.

Eric Keller, Colorado University

Abstract: Wireless networks face a variety of difficult present-day and future challenges, in particular those of burgeoning demands for capacity and the high costs and slow deployment of new wireless capabilities. To overcome the current state where there are multiple, provider specific static networks, (some even making use of the cloud), we believe we should learn from the successes of cloud computing, which made it trivial and low cost to roll out a new web service. We propose that industry should be doing the same with the wireless edge, in an architecture we call CloudBase. CloudBase uses software radio to enable a single device to be used for any wireless enabled service, virtualizes the software radio to enable a single device to be shared across multiple services, and organizes the virtualized software radios in a cloud-like architecture to enable dynamically deployable wireless services on a shared infrastructure.

Prof. Brighten Gdfrey (opens in new tab), University of Illinois at Urbana Champaign

Abstract: The increasing complexity of modern computer networks has far outpaced the development of tools to manage their operation. We are developing tools which simplify network security and management by formally reasoning about network-wide forwarding behavior. Our first data plane verification system, Anteater [SIGCOMM’11], revealed multiple real-world bugs in a large university network, including forwarding loops and stale ACL rules. VeriFlow [HotSDN’12] checks network-wide invariants in real time as each forwarding rule is inserted, optionally blocking vulnerabilities from being introduced into the network. Our current OpenFlow-based implementation can perform rigorous checking within hundreds of microseconds per rule insertion. This talk presents work with Ahmed Khurshid, Haohui Mai, Kelvin Zou, Wenxuan Zhou, Rachit Agarwal, Matthew Caesar, and Sam King.

Dr. Vijay Erramilli (opens in new tab), Telefonica

Abstract: In this short presentation, I’ll try to highlight 3 ongoing projects at Telefonica Research. The first project involves high-end cuisine and working with arguably, the best chef on the planet – Ferran Adria. The second project involves sports and working with arguably, the best football club on the planet — FC Barcelona. The last project is relatively sedate and deals with personal data and online privacy.

Dr. Atul Adya (opens in new tab), Google

Abstract: Over the past few years, we have built a large-scale notification system called Thialfi that provides notifications to a variety of Google applications such as Chrome, Google Drive, and Google Plus.

In this presentation, I will briefly describe Thialfi and share some of the lessons learnt working with various application teams and running Thialfi as a service. I believe that most of our insights are relevant for many large-scale services.

Dr. Sampath Rangarajan, NEC Labs America

Abstract: Mobile data traffic is expected to grow at a rapid pace. Cisco VNI report has predicted that mobile traffic will grow 18 times during the period 2011 – 2016. Although traditional PHY/MAC techniques for improving spectral efficiency is expected to provide some of the required increase in capacity, most of it is expected to come from new network-level techniques and architectural deployments. In this talk, I will introduce some of these technologies and on-going work at NEC Labs America in this direction.