Filter by topic and date
Banishing the bane of bufferbloat
- Bjørn Ivar Teigen IETF Participant
23 May 2023
Bufferbloat: It's a term that may not be well-known outside of technical circles, but it's a problem that affects everyone who uses the Internet. Bufferbloat refers to the excessive buffering of data packets in network routers and switches, which leads to increased latency. This results in frustratingly slow web browsing, laggy video calls, and overall poor quality of experience for Internet users.
![Shibuya Scramble](../../media/images/Scramble_from_above_SHIBUYA_SKY_49367160502_cro.original.jpg)
But fear not! The Internet Engineering Task Force (IETF) is actively working to combat bufferbloat and improve the performance of the Internet for everyone.
One of the key areas of focus for the IETF is congestion control, which involves managing the flow of data across networks to mitigate congestion. Less congestion means shorter queues, and thereby reduced latency. Congestion control is a critical aspect of network performance, and the IETF has been actively developing and standardizing congestion control algorithms to address bufferbloat.
One notable approach that the IETF has been working on is called Low Latency, Low Loss, Scalable Throughput (L4S). L4S is an innovative solution that combines the benefits of traditional congestion control with more detailed feedback from the network. By leveraging modern techniques used by Data Center TCP, L4S aims to reduce both latency and bufferbloat, resulting in a smoother and more responsive Internet experience.
The IETF has published several RFCs related to L4S, including RFC 9330, which outlines the architecture of L4S, RFC 9331, which defines the Explicit Congestion Notification (ECN) protocol for L4S, and RFC 9332, which describes the Dual-Queue Coupled Active Queue Management (AQM) algorithm for L4S. These RFCs provide technical details and guidance for implementing L4S in networks.
Furthermore, the IETF has also published RFC 8289 and RFC 8290, which define Controlled Delay Active Queue Management (CoDel) and Flow Queue CoDel (FQ-CoDel), respectively. These algorithms are designed to proactively manage network queues and prevent bufferbloat by controlling the delay and dropping of packets, leading to improved network and application performance. While congestion control works on a timescale measured in hundreds of milliseconds, FQ manages traffic on a smaller time-scale. Therefore, innovations like L4S and FQ-CoDel complement each other and should be deployed together. (FQ)-CoDel was originally developed by an open source community of academics and Linux developers, and later standardized in the IETF.
In addition to congestion control, the IETF is also actively involved in the development of measurement tools and techniques to better understand and quantify latency and, thereby, application performance in networks.
Other ongoing work in the IETF has focused on measurement methodologies, such as the "Responsiveness" Internet-Draft contributed by individuals working at Apple and the University of Cincinnati, the "Capacity Protocol" Internet-Draft by individuals from AT&T, and the "Precision Availability Metrics" Internet-Draft by IETF participants working at Ericsson, ZTE, Futurewei, and Inria. These efforts are crucial in helping network operators identify and address latency issues that can contribute to bufferbloat. Moving us towards smoother Internet experiences
The IETF's work on battling Internet latency in all its forms has far-reaching implications for the future of the Internet. As online applications continue to grow in complexity and demand, it becomes increasingly important to optimize the performance of networks to provide a seamless user experience. By developing congestion control algorithms like L4S, improving measurement techniques, and defining proactive queue management algorithms like CoDel and Flow Queue CoDel, the IETF is at the forefront of efforts to make the Internet smoother and more reliable for applications and users.
So, what does all of this mean for you as an Internet user? It means that you are likely already benefiting from the IETF's work without even realizing it, as reducing Internet latency contributes to a better online experience for everyone. IETF participants are actively working behind the scenes to develop technologies that will improve the performance of the Internet and tackle issues like bufferbloat. A wide range of companies are already implementing and deploying the results. The work of the IETF may not be visible to the average user, but it plays a crucial role in shaping the future of the Internet.
So the next time you enjoy a smooth video call, a fast download, or a lag-free online game, you can thank the IETF—and all of the people contributing their work—for their dedicated efforts in making the Internet better for all of us.
Curious about the quality of your Internet connection? Try one of_these_ tests for measuring bufferbloat and network responsiveness. Look for the latency or responsiveness measured during upload and download:
Developers looking for a tool to monitor network quality from inside an app can check out this free and open source tool.
Bjørn Ivar Teigen is Head of Research at Domos.