USENIX ATC '10 Technical Program

USENIX ATC '10 Banner

TECHNICAL PROGRAM

Just Up! Videos of the presentations are now available. Access is currently restricted to USENIX members and USENIX ATC '10 conference attendees. Not a member? Join today!

All sessions will take place in Back Bay C unless otherwise noted.

Check back here for updates to the schedule.

Session papers are available to conference registrants immediately and to everyone beginning June 23, 2010.

Proceedings Front Matter files are available to everyone: Cover, Copyright, ISBN | Title Page, Organizers, Reviewers | Table of Contents | Message from the Program Co-Chairs

Complete Proceedings are available only to registered attendees until the opening day of the technical sessions; they then become available to everyone.

NEW! E-Book Proceedings: Read the proceedings on the go in iPad-friendly EPUB format or Kindle-friendly Mobipocket format.

Tech Sessions:
Wednesday, June 23 | Thursday, June 24 | Friday, June 25 | Invited Speakers

Wednesday, June 23, 2010

8:00 a.m.–8:30 a.m. Continental Breakfast: Served in Republic A

8:30 a.m.–10:00 a.m.

Welcome, Awards, and Keynote Address

Back Bay C/D

KEYNOTE ADDRESS

Lessons of Scale at Facebook
Speaker: Bobby Johnson, Director of Engineering, Facebook, Inc.

Listen in MP3 format

In just over six years Facebook has grown from an idea in a dorm room to one of the most visited sites on the Internet. This explosive growth has created enormous technical challenges—we've had to quickly build systems to handle a doubling of load every six months to a year, and these systems are constantly changing. One of the reasons for our massive growth is our ability to quickly launch products and react to changes, but this is increasingly difficult as we face new constraints involved in running at massive scale.

I'll be talking about some specific technical challenges we've faced and the general principles we employ when addressing problems of scale. I'll also discuss how we structure our engineering process and culture to stay on top of unceasing growth while still moving fast to build new products.

10:00 a.m.–10:30 a.m. Break: Refreshments served in Republic A

10:30 a.m.–Noon

Session Chair: Andrew Baumann, ETH Zürich

DEFCON: High-Performance Event Processing with Information Security
Matteo Migliavacca and Ioannis Papagiannis, Imperial College London; David M. Eyers, University of Cambridge; Brian Shand, CBCU, Eastern Cancer Registry, National Health Service UK; Jean Bacon, Computer Laboratory, University of Cambridge; Peter Pietzuch, Imperial College London

Read the Abstract | Full paper | Slides

Wide-Area Route Control for Distributed Services
Vytautas Valancius and Nick Feamster, Georgia Institute of Technology; Jennifer Rexford, Princeton University; Akihiro Nakao, The University of Tokyo

Read the Abstract | Full paper | Slides

Awarded Best Paper!
LiteGreen: Saving Energy in Networked Desktops Using Virtualization
Tathagata Das, Microsoft Research India; Pradeep Padala, DOCOMO USA Labs; Venkat Padmanabhan and Ram Ramjee, Microsoft Research India; Kang G. Shin, The University of Michigan

Read the Abstract | Full paper

Noon–1:00 p.m. Lunch: Served in Republic AB

1:00 p.m.–2:00 p.m.

INVITED TALK

Visualizing Data
Ben Fry, Author and Consultant

The ability to collect and store data continues to increase, but our ability to understand it remains unchanged. In an attempt to gain better understanding of data, fields such as information visualization, data mining, and graphic design are employed, each solving an isolated part of the specific problem, but failing in a broader sense. I seek to bring the individual fields together as part of a single process. We'll explore Processing, an open source programming language and environment for images, animation, and interactions. It is used by tens of thousands of students, practitioners, artists, designers, researchers, and hobbyists for learning, prototyping, and production. I'll describe how I used Processing in my own work to rapidly prototype and refine complex data visualization experiments.

2:00 p.m.–3:00 p.m.

Session Chair: Nickolai Zeldovich, Massachusetts Institute of Technology

Stout: An Adaptive Interface to Scalable Cloud Storage
John C. McCullough, University of California, San Diego; John Dunagan and Alec Wolman, Microsoft Research, Redmond; Alex C. Snoeren, University of California, San Diego

Read the Abstract | Full paper | Slides

IsoStack—Highly Efficient Network Processing on Dedicated Cores
Leah Shalev, Julian Satran, Eran Borovik, and Muli Ben-Yehuda, IBM Research—Haifa

Read the Abstract | Full paper | Slides

3:00 p.m.–3:30 p.m. Break: Refreshments served in Republic A

3:30 p.m.–5:30 p.m.

Session Chair: Catherine Zhang, IBM T.J. Watson Research Center

A Realistic Evaluation of Memory Hardware Errors and Software System Susceptibility
Xin Li, Michael C. Huang, and Kai Shen, University of Rochester; Lingkun Chu, Ask.com

Read the Abstract | Full paper | Slides

The Utility Coprocessor: Massively Parallel Computation from the Coffee Shop
John R. Douceur, Jeremy Elson, Jon Howell, and Jacob R. Lorch, Microsoft Research

Read the Abstract | Full paper

Apiary: Easy-to-Use Desktop Application Fault Containment on Commodity Operating Systems
Shaya Potter and Jason Nieh, Columbia University

Read the Abstract | Full paper | Slides

Tolerating Malicious Device Drivers in Linux
Silas Boyd-Wickizer and Nickolai Zeldovich, MIT CSAIL

Read the Abstract | Full paper | Slides

6:00 p.m.–8:00 p.m.

Poster Session and Happy Hour

Republic AB

See the list of accepted posters here.

Thursday, June 24, 2010

9:00 a.m.–10:00 a.m.

KEYNOTE ADDRESS

Some Thoughts About Concurrency
Ivan Sutherland, Visiting Scientist at Portland State University

Our industry has grown up with a sequential model of computing, evolved to husband the logic associated with a few vacuum tubes. Now we must struggle to harness the vast concurrency of modern transistor circuits. Is concurrency fundamentally hard, or does it just seem hard because of our history of sequential programming? I believe some of each. Concurrency is fundamentally hard for only two reasons. One is that concurrent action requires coordination. The other is that concurrent action of many processes can produce an exponential explosion of states. How can we be sure that all such states are benign?

Concurrency is easy when we escape its details. Maybe instead of "programming sequential processes" we might better "configure concurrent communication." A communication view of computing matches well the cost structure of modern hardware, where logic is now essentially free but moving data is relatively slow and expensive in time and energy. Making communication central to computation also prepares us for the increasing role geometry will play in the future of computing. New thinking may be essential to harnessing the vast concurrency provided by modern transistor circuits.

10:00 a.m.–10:30 a.m. Break: Refreshments served in Republic A

10:30 a.m.–Noon

Session Chair: Michael Stumm, University of Toronto

Proxychain: Developing a Robust and Efficient Authentication Infrastructure for Carrier-Scale VoIP Networks
Italo Dacosta and Patrick Traynor, Converging Infrastructure Security (CISEC) Laboratory, Georgia Tech Information Security Center (GTISC), Georgia Institute of Technology

Read the Abstract | Full paper | Slides

Awarded Best Paper!
ZooKeeper: Wait-free Coordination for Internet-scale Systems
Patrick Hunt and Mahadev Konar, Yahoo! Grid; Flavio P. Junqueira and Benjamin Reed, Yahoo! Research

Read the Abstract | Full paper | Slides

Testing Closed-Source Binary Device Drivers with DDT
Volodymyr Kuznetsov, Vitaly Chipounov, and George Candea, École Polytechnique Fédérale de Lausanne (EPFL), Switzerland

Read the Abstract | Full paper | Slides

Noon–1:00 p.m. Lunch: Served in Republic AB

1:00 p.m.–3:00 p.m.

Session Chair: Steven Hand, University of Cambridge and Citrix Systems

A Transparently-Scalable Metadata Service for the Ursa Minor Storage System
Shafeeq Sinnamohideen and Raja R. Sambasivan, Carnegie Mellon University; James Hendricks, Carnegie Mellon University and Google; Likun Liu, Tsinghua University; Gregory R. Ganger, Carnegie Mellon University

Read the Abstract | Full paper | Slides

FlashVM: Virtual Memory Management on Flash
Mohit Saxena and Michael M. Swift, University of Wisconsin—Madison

Read the Abstract | Full paper | Slides

Dyson: An Architecture for Extensible Wireless LANs
Rohan Murty, Harvard University; Jitendra Padhye and Alec Wolman, Microsoft Research; Matt Welsh, Harvard University

Read the Abstract | Full paper

ChunkStash: Speeding Up Inline Storage Deduplication Using Flash Memory
Biplob Debnath, University of Minnesota, Twin Cities; Sudipta Sengupta and Jin Li, Microsoft Research, Redmond

Read the Abstract | Full paper | Slides

3:00 p.m.–3:30 p.m. Break: Refreshments served in Republic A

3:30 p.m.–4:30 p.m.

INVITED TALK

Google Books: Making All the World's Books Universally Accessible and Useful
Jon Orwant, Engineering Manager, Google

Google has now digitized a significant percentage of all the books ever published, and our corpus now weighs in at two trillion words. We've overcome a number of challenges along the way, and in this talk I'll describe some of them: scanning, image processing, storage, serving, structure extraction, and the seemingly simple problem of identifying which books exist.

4:30 p.m.–6:00 p.m.

Work-in-Progress Reports (WiPs)

See the list of accepted WiPs here.

6:30 p.m.–8:00 p.m.

Reception

Republic AB

Friday, June 25, 2010

9:00 a.m.–10:00 a.m.

INVITED TALK

Reconstructing Ancient Rome: 700 Years of IT and Knowledge Management
Maximilian Schich, DFG Visiting Research Scientist at BarabásiLab, Center for Complex Network Research at Northeastern University

Documentation and reconstruction of ancient Rome has been an IT and knowledge management problem ever since the 15th century. First we will examine the fascinating phenomenon of sample and remix in building reconstruction which emerged firmly in the 16th century and, surprisingly, still continues today, even if present-day archaeologists are not aware of it. Then we will see how modern art historians learned to cope with ever-increasing amounts of documentation as they moved from the production of encyclopedic books to card indices and sophisticated image databases. Finally we will see how these datasets of documents become subjects of inquiry themselves. Using network analysis and visualization, we can now easily map the contents of entire datasets, allowing for new insights, big pictures, comparison, and project evaluation.

10:00 a.m.–10:20 a.m. Break: Refreshments served in Republic A

10:20 a.m.–Noon

Session Chair: Galen Hunt, Microsoft Research Redmond

Sleepless in Seattle No Longer
Joshua Reich, Columbia University; Michel Goraczko, Aman Kansal, and Jitendra Padhye, Microsoft Research

Read the Abstract | Full paper | Slides

Wide-area Network Acceleration for the Developing World
Sunghwan Ihm, Princeton University; KyoungSoo Park, University of Pittsburgh and KAIST; Vivek S. Pai, Princeton University

Read the Abstract | Full paper | Slides

An Evaluation of Per-Chip Nonuniform Frequency Scaling on Multicores (Short Paper)
Xiao Zhang, Kai Shen, Sandhya Dwarkadas, and Rongrong Zhong, University of Rochester

Read the Abstract | Full paper | Slides

A DNS Reflection Method for Global Traffic Management (Short Paper)
Cheng Huang, Microsoft Research; Nick Holt, Microsoft Corporation; Y. Angela Wang, Polytechnic Institute of NYU; Albert Greenberg and Jin Li, Microsoft Research; Keith W. Ross, Polytechnic Institute of NYU

Read the Abstract | Full paper | Slides

Noon–1:00 p.m. Lunch: Served in Republic AB

1:00 p.m.–2:00 p.m.

INVITED TALK

RoboBees: An Autonomous Colony of Robotic Pollinators
Matt Welsh, Associate Professor of Computer Science at Harvard University

Our team of researchers from Harvard, Northeastern, and Centeye is developing a coordinated colony of robotic bees, aptly called RoboBees. The idea is to build a swarm of flapping-wing microrobots that are capable of performing wide-area search, environmental mapping and surveillance, and even pollination. This project involves a wide range of research challenges across many fields, broken down into three categories: body, brain, and colony. Problems in the "body" space include all aspects of the flight apparatus, propulsion, and power supply. The "brain" involves research on low-power electronics, including circuits for sensing and decision-making. Finally, research within the "colony" entails communication and control algorithms that will enable performance well beyond the capabilities of an individual RoboBee. Real bees coordinate to interact with the complex environment using a diversity of sensors, a hierarchy of task delegation, unique communication, and an effective flapping-wing propulsion system. Mimicking these tasks involves many open research problems that drive this project. In this talk I will present the RoboBees vision, the core technologies we are building upon, and our progress to date.

View the presentation slides

2:00 p.m.–3:00 p.m.

Session Chair: Dave Presotto, Google, Inc.

An Analysis of Power Consumption in a Smartphone
Aaron Carroll, NICTA and University of New South Wales; Gernot Heiser, NICTA, University of New South Wales, and Open Kernel Labs

Read the Abstract | Full paper | Slides

SleepServer: A Software-Only Approach for Reducing the Energy Consumption of PCs within Enterprise Environments
Yuvraj Agarwal, Stefan Savage, and Rajesh Gupta, University of California, San Diego

Read the Abstract | Full paper

3:00 p.m.–3:30 p.m. Break: Refreshments served in Republic A

3:30 p.m.–4:30 p.m.

Session Chair: Timothy Roscoe, ETH Zürich

An Extensible Technique for High-Precision Testing of Recovery Code
Paul D. Marinescu, Radu Banabic, and George Candea, École Polytechnique Fédérale de Lausanne (EPFL), Switzerland

Read the Abstract | Full paper

Mining Invariants from Console Logs for System Problem Detection
Jian-Guang Lou and Qiang Fu, Microsoft Research Asia; Shenqi Yang, Beijing University of Posts and Telecom; Ye Xu, Nanjing University, P.R. China; Jiang Li, Microsoft Research Asia

Read the Abstract | Full paper | Slides

Need help? Use our Contacts page.

Last changed: 14 Sept. 2010 ch