Internet Congestion Control (iccrg)
|Internet Congestion Control
|Reese Enghardt, Simone Ferlin, Vidhi Goel
Charter for Research Group
Internet congestion control crucially impacts network efficiency and user
quality of experience (QoE). The Internet Congestion Control Research Group
(ICCRG) is an IRTF group tasked with investigating innovative conceptual
and practical aspects of Internet congestion control. The ICCRG brings
together researchers and practitioners from academia and industry and
strives to promote continued engagement by participants to facilitate an
ongoing and lively discussion of congestion control research. Discussions
in this group range from early, half-baked ideas that could benefit from
feedback from the community to evaluations of widely-deployed congestion
controllers and their impact on the Internet.
Topics of Interest:
The ICCRG’s primary focus is on the open Internet, broadly defined, though
its interests extend to related network domains. In particular, topics of
interest for the ICCRG include:
1) Trends impacting Internet congestion control:
- trends in application requirements: different types of emerging /
existing application profiles, and the induced requirements from
the network needed to support high QoE;
- trends in traffic patterns: different types of emerging/existing
traffic profiles and how to best support them;
- interactions with the network: specific network characteristics
(e.g., 5G, LTE, wired), QoS mechanisms, signaling mechanisms
(e.g., ECN, INT), traffic engineering, lower-layer technologies;
- interactions with other traffic: traffic controlled by different
congestion controllers, DDoS and mitigation mechanisms, other
forms of bandwidth consumption/protection.
2) Novel approaches to Internet congestion control protocol design:
- new design objectives, e.g., application-oriented/QoE-aware protocol
- new algorithmic approaches, e.g., control-theoretic, learning-based
- rethinking the design space, e.g., novel schemes for multipath
3) New evaluation frameworks for congestion control.
- new formal models and performance metrics/criteria for congestion
control, spanning different notions of performance, fairness,
- network simulation/emulation frameworks for congestion control,
including new experimentation tools and proposed experiments.
- empirical datasets for evaluating congestion control protocols that
reflect real-world challenges for Internet congestion control.
To accommodate reproducibility, we strongly encourage participants to share
code for congestion control prototypes and evaluation tools, as well as
(possibly anonymized) useful datasets with the community.
4) Evaluation of (existing and new) congestion controllers.
- Theoretical results for Internet congestion control, e.g., theoretical
guarantees for specific congestion controllers, general impossibility
results, inherent tradeoffs between different objectives, etc.
- Experimental/empirical evaluation results describing the behavior of
different congestion controllers and contrasting different congestion
control schemes. This includes comparative academic studies, as well
as input from practitioners working on congestion control, even if
limited to their own network environments or use cases.
If you are working on something that you believe is in scope for ICCRG but
are uncertain, please reach out to the chairs.
The ICCRG’s goal is to enhance our understanding of Internet congestion
control and to support conceptual innovation. Hence, while ICCRG
discussions might entail writing documents, these are not a required or
expected outcome of the ICCRG. In particular, the ICCRG does not produce
standards-track RFCs (although it can publish Experimental or Informational
RFCs). The IETF is the appropriate venue for deployment-oriented documents
concerning interoperability or standardization of implementation, where the
ICCRG will coordinate with the IETF transport area as appropriate.
While oral presentations of ideas at ICCRG meetings and any subsequent
discussions are the primary form of communicating ideas at the ICCRG,
participants are encouraged to deposit artifacts, such as datasets, draft
research papers, documents covering recent theoretical/empirical results on
Internet congestion control, for review, discussion, or collaboration.
These artifacts are not intended as replacements to Internet drafts and may
be provided in any appropriate format.
The ICCRG will coordinate with other IRTF research groups, the IETF
Transport Area, practitioners in industry, and the broader research
community as appropriate.