TCP’s Initial Window—Deployment in the Wild and Its Impact on Performance

Rüth, Jan; Kunze, Ike; Hohlfeld, Oliver

doi:10.1109/tnsm.2019.2896335

Cited by 7 publications

(7 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An AR of 10 was chosen as a starting point given the mounting evidence that Internet transport protocols can sustain an initial burst of 10 packets. 8 Here we extend this work by examining the effect of an AR of 20 and 100, on a variety of network paths. We also explore how the AR can influence throughput at the start of a QUIC connection and in steady-state, with and without packet loss.…”

mentioning

confidence: 81%

“…This paper builds on previous work 7 by the authors, where AR 10 was evaluated in GEO satellite networks. An AR of 10 was chosen as a starting point given the mounting evidence that Internet transport protocols can sustain an initial burst of 10 packets 8 . Here we extend this work by examining the effect of an AR of 20 and 100, on a variety of network paths.…”

Section: Introductionmentioning

confidence: 92%

See 1 more Smart Citation

Reducing the acknowledgement frequency in IETF QUIC

Custura

Jones

Secchi

et al. 2022

Satell Commun Network

View full text Add to dashboard Cite

Satellite networks usually use in-network methods (such as Performance Enhancing Proxies for TCP) to adapt the transport to the characteristics of the forward and return paths. QUIC is a transport protocol that prevents the use of in-network methods. This paper explores the use of the recently-standardised IETF QUIC protocol with a focus on the implications on performance when using different acknowledgement policies to reduce the number of packets and volume of bytes sent on the return path. Our analysis evaluates a set of ACK policies for three IETF QUIC implementations, examining performance over cellular, terrestrial and satellite networks. It shows that QUIC performance can be maintained even when sending fewer acknowledgements, and recommends a new QUIC acknowledgement policy that adapts QUIC's ACK Delay value based on the path RTT to ensure timely feedback.The resulting policy is shown to reduce the volume/rate of traffic sent on the return path and associated processing costs in endpoints, without sacrificing throughput.

show abstract

mentioning

confidence: 81%

Section: Introductionmentioning

confidence: 92%

Reducing the acknowledgement frequency in IETF QUIC

Custura

Jones

Secchi

et al. 2022

Satell Commun Network

View full text Add to dashboard Cite

show abstract

“…RTO can indeed affect TCP performance in the wild. At the same time, in [11], the authors analyzed the TCP's initial window in the wild, also in this case with a massive real test campaign.…”

Section: Related Workmentioning

confidence: 99%

Future of TCP on Wi-Fi 6

Grazia

2021

IEEE Access

View full text Add to dashboard Cite

The Linux TCP/IP stack contributions have recently pointed all in the same direction: maximize the available throughput while maintaining low latency. These activities started by mitigating the buffebloat phenomenon at the network bottleneck as much as possible. So far, the deployed solutions have been designed by considering standard models for the bottleneck that could be either wireless or wired. The introduction of Wi-Fi 6 is dislocating the bottleneck of standard home and office WLAN from the radio access point to the 1 Gbps interface, the one that points to the internet service provider; this bottleneck migration leads to a new real-case bottleneck model, which is hybrid. Such an environment embraces new technologies and provides new challenges for the old TCP protocol when applied to hybrid bottlenecks and operating in conjunction with the TCP side-modules, which are now part of the novel Linux kernels. This paper aims to highlight the TCP performance considering the new TCP modules and novel scenarios opened by Wi-Fi 6 with real-case hybrid bottlenecks.

show abstract

“…One line of research aims at characterizing remote TCP stacks by their behavior (e.g., realized in the TCP Behavior Inference Tool (TBIT) [45] in 2001). One aspect is to study the deployment of TCP tunings (e.g., the initial window configuration [48][49][50]) or TCP extensions (e.g., Fast Retransmit [45], Fast Open [39,44], Selective Acknowledgment (SACK) [36,43,45], or Explicit Congestion Notification (ECN) [18,36,37,42,43,45] and ECN++ [38] to name a few). While these works aim to generally characterize stacks by behavior and to study the availability and deployability of TCP extensions, our work specifically focuses on the conformance of current TCP stacks to mandatory behavior every stack must implement.…”

Section: Related Workmentioning

confidence: 99%

MUST, SHOULD, DON’T CARE: TCP Conformance in the Wild

Kosek

Blöcher

Rüth

et al. 2020

Passive and Active Measurement

Self Cite

View full text Add to dashboard Cite

Standards govern the SHOULD and MUST requirements for protocol implementers for interoperability. In case of TCP that carries the bulk of the Internets' traffic, these requirements are defined in RFCs. While it is known that not all optional features are implemented and nonconformance exists, one would assume that TCP implementations at least conform to the minimum set of MUST requirements. In this paper, we use Internet-wide scans to show how Internet hosts and paths conform to these basic requirements. We uncover a non-negligible set of hosts and paths that do not adhere to even basic requirements. For example, we observe hosts that do not correctly handle checksums and cases of middlebox interference for TCP options. We identify hosts that drop packets when the urgent pointer is set or simply crash. Our publicly available results highlight that conformance to even fundamental protocol requirements should not be taken for granted but instead checked regularly.

show abstract

TCP’s Initial Window—Deployment in the Wild and Its Impact on Performance

Cited by 7 publications

References 18 publications

Reducing the acknowledgement frequency in IETF QUIC

Reducing the acknowledgement frequency in IETF QUIC

Future of TCP on Wi-Fi 6

MUST, SHOULD, DON’T CARE: TCP Conformance in the Wild

Contact Info

Product

Resources

About