Towards a data-driven framework for optimizing security-efficiency tradeoff in QUIC

Murthy, Amith; Asghar, Muhammad Rizwan; Tu, Wanqing

Abstract

Advances in computing and compression technology, coupled with high-speed networks, has beaconed an era of video streaming on the Internet. This has led to a need to enhance the security of communications transporting data without degrading its performance. The transport layer security (TLS) protocol negotiates configurations for securing communication channels. Such conversations adversely impact latency, thereby presenting a fundamental tradeoff between security and efficiency. In this work, we present a conceptual framework, called SEC-QUIC (Secure and Efficient Configurations for QUIC), that focuses on optimizing this tradeoff specifically for video transmissions by investigating various factors in Quick UDP Internet Connections (QUIC). Transport-layer-related elements, such as maximum transmission unit (MTU) sizes, cipher suites, and ACK timer, are examined to evaluate the impact on the security-efficiency tradeoff in QUIC-based video transmissions using platform-based experiments. Subsequently, we develop a conceptual framework to leverage QUIC's dynamics based on the context of a connection to optimize the security-efficiency tradeoff. Our findings demonstrate the need to alter default configurations based on the contextual factors of a connection (eg, resource constraints and network conditions) in QUIC-based video transmissions to balance the tradeoff. Experiments reveal an MTU of 1400 bytes is found to have 60% better throughput compared to an MTU of 1200 bytes while also 4% less CPU usage on average for the transmission of 100 MB video files. Overall, our experiments suggest that fine-tuning performance and security related configurations is an effective approach to optimizing the security-efficiency tradeoff in video transmissions.

Citation

Murthy, A., Asghar, M. R., & Tu, W. (2022). Towards a data-driven framework for optimizing security-efficiency tradeoff in QUIC. Security and Privacy, 5(1), Article e184. https://doi.org/10.1002/spy2.184

• Journal Article Type: Article
• Acceptance Date: Jul 27, 2021
• Online Publication Date: Aug 11, 2021
• Publication Date: Jan 3, 2022
• Deposit Date: Oct 27, 2021
• Publicly Available Date: Aug 11, 2022
• Journal: Security and Privacy
• Electronic ISSN: 2475-6725
• Publisher: Wiley
• Peer Reviewed: Peer Reviewed
• Volume: 5
• Issue: 1
• Article Number: e184
• DOI: https://doi.org/10.1002/spy2.184
• Public URL: https://durham-repository.worktribe.com/output/1232934