QUIC: Does the Protocol Achieve its Goals.

Austin Benoit, Chris Holland

Problem Statement

QUIC (Quick UDP Internet Connections) is a new transport layer protocol created by Google which aims to reduce the latency of that of TCP. QUIC is built on UDP, and therefore is not as difficult to make changes to compared to TCP; since TCP is implemented in operating system kernels and middlebox firmware. Google has recently built an undersea network cable and the Google CDN (Content Delivery Network); and Chrome has been the most popular web browser for years. This means that Google has dominion over the Link, Network, and Application layers of the interent, however, not the Transport Layer. This is where QUIC comes in [1] [2].

Over the course of this semester we will analyze the QUIC protocol to determine whether it’s architecture and implementation are sufficient in meeting the goals it sets out to achieve (e.g. lower latency times compared to TCP). The goal will be to build a deeper understanding of the protocol, learn about how it is implemented, and discover if anything can be improved in order to better meet its business goals.

Previous Work

This project will build upon the work done by Henry Baxter and Liam Simmons in an earlier offering of the CSC 466 course [3]. In their project they configured virtual machines to analyze the latency of QUIC and TCP and measure analytics. We aim to use their data and discover why the QUIC protocol performs the way it does through analysis of its implementation and architecture.

Analytical Approach

QUIC is described as a “multiplexed and secure general-purpose transport protocol” that provides:

• Stream multiplexing
• Stream and connection-level flow control
• Low-latency connection establishment
• Connection migration and resilience to NAT rebinding
• Authenticated and encrypted header and payload

QUIC uses UDP as a substrate to avoid requiring changes to legacy client operating systems and middleboxes [4]. Given this information, we expect to follow the approach below in order to complete our research on QUIC.

• Review the current literature/documentation about QUIC
• Study the implementation and code surrounding QUIC in current form
• Run tests on the key points of the QUIC protocol
• Compile the results of tests and tie in any relevant literature
• Use the compiled results and compare them with goals outlined in QUIC as stated above

Expected Timeline

Week of	Deliverables
Jan 27	Project Proposal
Feb 3	Review current QUIC implementation and architecture
Feb 17	Run analytical tests
March 2	Explore findings
March 16	Create report
March 23	Prepare for presentation. Finalize the report

References

1. https://www.chromium.org/quic
2. https://docs.google.com/presentation/d/15e1bLKYeN56GL1oTJSF9OZiUsI-rcxisLo9dEyDkWQs/edit#slide=id.g5b4208ab1_0_311
3. https://csc466quic.wordpress.com/
4. https://tools.ietf.org/html/draft-ietf-quic-transport-25