[Paper Review] A Fresh Look at ECN Traversal in the Wild

Hyoyoung Lim, Seonwoo Kim+8 more|arXiv (Cornell University)|Aug 30, 2022

Image and Video Quality Assessment4 citations

TL;DR

This paper presents a large-scale active and passive measurement study of ECN traversal across wired and cellular networks, identifying and localizing ECN bleaching—where intermediate devices clear ECN bits—across 129,252 routers. It finds that 4.17% of paths exhibit ECN bleaching, with some providers disabling ECN by policy, undermining ECN and L4S deployment for low-latency applications.

ABSTRACT

The Explicit Congestion Notification (ECN) field has taken on new importance due to Low Latency, Low Loss, and Scalable throughput (L4S) technology designed for extremely latency-sensitive applications (such as cloud games and cloud-rendered VR/AR). ECN and L4S need to be supported by the client and server but also all devices in the network path. We have identified that "ECN bleaching", where an intermediate network device clears or "bleaches" the ECN flags, occurs and quantified how often that happens, why it happens and identified where in the network it happens. In this research, we conduct a comprehensive measurement study on end-to-end traversal of the ECN field using probes deployed on the Internet across different varied clients and servers. Using these probes, we identify and locate instances of ECN bleaching on various network paths on the Internet. In our six months of measurements, conducted in late 2021 and early 2022, we found the prevalence varied considerably from network to network. One cloud provider and two cellular providers bleach the ECN field as a matter of policy. Of the rest, we found 1,112 out of 129,252 routers, 4.17% of paths we measured showed ECN bleaching.

Motivation & Objective

• To measure the end-to-end usability of ECN in real-world internet paths, especially across diverse wired and cellular networks.
• To identify and localize instances of ECN bleaching—where intermediate devices clear ECN bits—across the network path.
• To quantify the prevalence of ECN bleaching and determine which network providers or devices are responsible.
• To assess the impact of ECN bleaching on the deployment of L4S, a new congestion control standard for low-latency applications.
• To provide actionable data to ISPs and protocol designers to improve ECN interoperability and deployment.

Proposed method

• Deployed active probing tools across diverse clients and servers to test end-to-end ECN traversal on real internet paths.
• Used a custom measurement tool to detect and locate ECN bleaching by analyzing changes in ECN field values across network hops.
• Conducted passive measurements on a university network to analyze ECN field usage in real traffic over a 24-hour period.
• Leveraged PATHspider to measure ECN capability on web servers and assess server-side ECN adoption.
• Combined active and passive measurements to validate bleaching detection and improve path localization accuracy.
• Shared findings with ISPs to elicit feedback and encourage remediation of ECN-bleaching devices.

Experimental results

Research questions

• RQ1 What is the prevalence of ECN bleaching across different types of internet networks, including wired and cellular providers?
• RQ2 Which network devices or ISPs are responsible for ECN bleaching, and is it due to policy or misconfiguration?
• RQ3 How does ECN bleaching affect the deployment and effectiveness of L4S, a new low-latency congestion control standard?
• RQ4 To what extent does ECN traversal remain end-to-end transparent in modern internet paths?
• RQ5 Are there geographical or infrastructure-based patterns in ECN bleaching, and can they be localized to specific network segments?

Key findings

• In a six-month measurement period, 1,112 out of 129,252 routers (4.17%) were found to bleach ECN bits, indicating widespread but non-universal ECN traversal failure.
• One cloud provider and two cellular providers explicitly bleach ECN as a matter of policy, significantly limiting ECN and L4S deployment.
• ECN bleaching is not uniformly distributed; it is more prevalent in certain network segments, particularly in mobile access networks.
• Despite widespread ECN support in modern operating systems, only 86.4% of web servers tested were ECN-capable, indicating ongoing deployment challenges.
• The study identified that ECN negotiation via TCP flags can succeed even when IP-level ECN is bleached, highlighting a potential workaround for some applications.
• Feedback from ISPs after sharing results showed willingness to address ECN bleaching, suggesting that the findings can drive real-world protocol improvements.