ACM SIGCOMM 2020 Workshop on Evolution, Performance, and Interoperability of QUIC (EPIQ 2020)
Workshop program
The workshop has an associated Slack channel for discussions. Click on the link below to visit it. If you're asked to sign in, use the workspace name "sigcomm.slack.com" to sign up or sign in.
Go to workshop Slack channel-
10:00 - 10:15 am EDT Chairs welcome
-
10:15 - 11:15 am EDT As QUIC as TCP, Optimizing QUIC and HTTP/3 CPU usage
Ian Swett (Google)
-
Abstract: TCP has benefitted from years of optimization. As recently as 2017, QUIC consumed twice the CPU of TLS over TCP. However, QUIC was designed as an encrypted transport, which benefits application performance and can benefit transport CPU performance. This talk will discuss CPU optimizations that have brought QUIC CPU usage down to parity with TCP in some cases, as well as looking at optimizations and hardware offloads that could enable QUIC to surpass TCP in CPU consumption.
Bio: Ian Swett is a software engineer who’s been working on QUIC at Google prior to its first public use and is an IETF editor on the loss detection and congestion control draft. Ian led the work to improve QUIC CPU usage at Google, with a focus on YouTube video delivery.
-
-
11:15 - 11:30 am EDT Short break
- Tea/Coffee/Meal Break
-
11:30 - 11:59 am EDT Testing QUIC with packetdrill
Vidhi Goel, Rui Paulo, and Christoph Paasch
-
Abstract: A secure transport protocol like QUIC has extensive state machines and algorithms to ensure secure, reliable and in-order delivery of multiple byte streams. Testing these mechanisms is at the heart of a successful deployment of QUIC. While interop-testing, fuzzing and performance tests allow to explore a large part of these algorithms, concrete issues are hard to reproduce reliably with those methods. packetdrill has emerged as an effective tool for exercising the TCP stack by generating events like incoming packets or system calls at precise timings.
In this paper, we present how we extended packetdrill for testing QUIC. This involves a special grammar to express the QUIC packet format and integration of TLS 1.3 into packetdrill. We integrated our own QUIC library and show that it is easy to include any QUIC library into packetdrill, and even reuse test cases across different libraries. This illustrates that using packetdrill is a useful approach for testing QUIC implementations.
-
-
12:00 - 12:30 pm EDT Automating QUIC Interoperability Testing
Marten Seemann and Jana Iyengar
-
Abstract: We present QuicInteropRunner (QIR) [1, 2], a test framework for automated and on-demand interoperability testing between implementations of the QUIC protocol [3]. QIR is a framework in which QUIC clients and servers interact with each other over a network that simulates various network conditions using ns-3 [4]. QIR automates QUIC interoperability testing by running a suite of test cases between containerized QUIC implementations. We describe the key constraints and insights that defined our work, recent innovations that made the framework possible, a high-level overview of our design, and a few exemplary tests. QIR is now supported and used by ten QUIC implementations as part of their development process, confirming our thesis that there is a need for automating interoperability testing and making it available on demand.
-
-
12:30 - 1:00 pm EDT Same Standards, Different Decisions: A Study of QUIC and HTTP/3 Implementation Diversity
Robin Marx, Joris Herbots, Wim Lamotte, and Peter Quax
-
Abstract: The QUIC and HTTP/3 protocols are quickly maturing together with their implementations, though many of their low-level behaviours are not yet well-understood. To help improve this, we empirically compare 15 IETF QUIC+HTTP/3 implementations for advanced features like Flow and Congestion Control, 0-RTT, Multiplexing, and Packetization. We find a large heterogeneity between stacks, discuss uncovered bugs and conclude that most implementations are not fully optimized or validated yet. We argue that future work must prioritize rigorous root-cause analysis of observed behaviours, and show this is possible by employing our qlog and qvis tools.
-
-
1:00 - 2:00 pm EDT Break
- Tea/Coffee/Meal Break
-
2:00 - 2:30 pm EDT Making QUIC Quicker With NIC Offload
Xiangrui Yang, Lars Eggert, Jörg Ott, Steve Uhlig, Zhigang Sun, and Gianni Antichi
-
Abstract: This paper aims at defining the right set of primitives a NIC shall expose to efficiently offload the QUIC protocol. Although previous work already partially tackled this problem, it has only considered one specific aspect: the crypto module. We instead dissect different QUIC implementations, and perform an in-depth analysis of the cost associated to many of its components. We find that the kernel to userspace communication, the crypto module and the packet reordering algorithm are CPU hungry and often the cause of application performance degradation. We use those findings to define an architecture for offloading QUIC and discuss the associated challenges.
-
-
2:30 - 3:00 pm EDT Scalable High Efficiency Video Coding based HTTP Adaptive Streaming over QUIC
Minh Nguyen, Hadi Amirpour, Christian Timmerer, and Hermann Hellwagner
-
Abstract: HTTP/2 has been explored widely for adaptive video streaming, but still suffers from Head-of-Line blocking, and three-way handshake delay due to TCP. Meanwhile, QUIC running on top of UDP can tackle these issues. In addition, although many adaptive bitrate (ABR) algorithms have been proposed for scalable and non-scalable video streaming, the literature lacks an algorithm designed for both types of video streaming approaches. In this paper, we investigate the impact of QUIC and HTTP/2 on the performance of ABR algorithms. Moreover, we propose an efficient approach for utilizing scalable video coding formats for adaptive video streaming that combines a traditional video streaming approach (based on non-scalable video coding formats) and a retransmission technique. The experimental results show that QUIC benefits significantly from our proposed method in the context of packet loss and retransmission. Compared to HTTP/2, it improves the average video quality and provides a smoother adaptation behavior. Finally, we demonstrate that our proposed method originally designed for non-scalable video codecs also works efficiently for scalable videos such as Scalable High Efficiency Video Coding (SHVC).
-
-
3:00 - 3:30 pm EDT Analyzing the Adoption of QUIC From a Mobile Development Perspective
Diego Madariaga, Lucas Torrealba, Javier Madariaga, Javiera Bermúdez, and Javier Bustos-Jiménez
-
Abstract: First introduced in 2013, QUIC protocol has been in constant development and it has gained great importance over time, currently being standardized by the Internet Engineering Task Force (IETF). Previous experiments have shown that the performance of QUIC can be of particular interest for lossy wireless networks. Thus, it becomes relevant to measure and profile QUIC traffic from mobile devices in the wild. This paper analyzes the adoption of QUIC from a mobile development perspective, by using crowdsourced data from real Android users during the normal use of their mobile devices. The results obtained show a significant increase in the number of Android apps using QUIC and evidence newer efforts of companies to adopt this protocol. To the best of the authors' knowledge, this is the first work focused on analyzing and profiling QUIC traffic from crowdsourced network measurements taken by mobile end-user devices.
-
-
3:30 - 3:45 pm EDT End
Call for Papers
The transport protocol QUIC has emerged from a proprietary effort undertaken by Google to a next generation transport protocol being standardized in the Internet Engineering Task Force (IETF). While its original motivation and design was to support next-generation Web traffic using HTTP/2, embedding QUIC into the Internet architecture raises exciting challenges beyond the necessary engineering efforts. QUIC has already seen deployment and motivated research papers measuring, extending, and evaluating QUIC from various perspectives.
The Second ACM SIGCOMM Workshop on the Evolution, Performance, and Interoperability of QUIC (EPIQ) seeks to continue to foster this emerging community. We invite researchers from academia and industry as well as engineers to explore novel ideas and future directions of QUIC and its interaction with applications and networks.
EPIQ solicits two types of submissions, for presentation and discussion at the workshop: academic papers and posters & demos. Submissions of both types should focus on topics related to the rise of QUIC on the mobile and fixed Internet as well as enterprise and datacenter networks. Papers focusing on either the original flavor of QUIC currently deployed by Google or the upcoming IETF standard of QUIC are in scope. We encourage submissions of demos supporting growing a community, e.g., about tools for protocol testing, deployment, and performance evaluation, among others. Open source tools are preferred.
Topics of Interest
Topics of interest include:- Qualitative and/or quantitative comparisons of QUIC to other protocols
- Tools for QUIC interoperability testing, validation, and conformance
- Formal models for and verification of the QUIC protocol architecture
- Measurements of QUIC traffic in the wild and profiling of QUIC implementations
- Acceleration techniques, including hardware offloading or specific OS improvements
- Scalable QUIC implementations and load balancing
- Advanced QUIC features, including integration of multicast or multipath connectivity
- Peer-to-peer QUIC and NAT/firewall traversal
- New abstractions and APIs for QUIC as a general-purpose transport protocol
- Troubleshooting, managing and monitoring of QUIC traffic in the network
- Privacy and security aspects of QUIC or of systems impacted by QUIC
- New applications for QUIC (e.g., real-time interactive)
- Novel approaches for congestion control enabled by QUIC
- QUIC as a substrate for non-HTTP applications
- Downscaling QUIC for IoT or embedded use
We encourage the submission of early results, negative results, and independently reproduced results of previously published work.
Important Dates
-
May 08, 2020 11:59 PST
Paper submission deadline
-
May 25, 2020 11:59 PST
Paper acceptance notification
-
June 10, 2020 11:59 PST
Camera-ready deadline
-
August 14, 2020
Workshop
Committees
- Program Committee Chairs
-
Jörg Ott
Technical University Munich
-
Lars Eggert
NetApp
- Technical Program Committee
-
Vaibhav Bajpai
TU Munich
-
Olivier Bonaventure
UC Louvain
-
Anna Brunström
Karlstad University
-
Stephen Farrell
Trinity College Dublin
-
Simone Ferlin
Ericsson
-
Oliver Hohlfeld
TU Cottbus
-
Michio Honda
University of Edinburgh
-
Jana Iyengar
Fastly
-
Subodh Iyengar
Facebook
-
Kathy Nichols
Pollere
-
Cristina Nita-Rotaru
Northeastern University
-
Colin Perkins
University of Glasgow
-
Ian Swett
Google
-
Michael Welzl
University of Oslo
-
Oliver Hohlfeld
Brandenburg University of Technology
Submission Instructions
EPIQ solicits two types of submissions, for presentation and discussion at the workshop: academic papers and posters & demos.
Paper submissions may be up to six pages in length, including all figures, tables, and appendices, but excluding references, for which up to one extra page may be used. The review process is single-blind. Papers must otherwise follow the formatting requirements of the main ACM SIGCOMM 2020 conference. Accepted papers will be published in the ACM Digital Library. Authors of accepted papers are required to present and discuss their papers in person at the workshop.
Poster & demo submissions may be up to two pages in length, and should describe the content of the poster or demo. Demo submissions must describe any venue requirements beyond WLAN connectivity and table space. Accepted poster & demo submissions will be highlighted in a very short “lightning talk” to advertise their appearance in a dedicated joint poster & demo session. Poster & demo submissions will not be included in the proceedings, but may be made available via the workshop website.
Please submit papers and poster&demo proposals at https://epiq20.hotcrp.com.
