RAPID: Avoiding TCP Incast Throughput Collapse in Public Clouds With Intelligent Packet Discarding

Xu, Yang; Shukla, Shikhar; Guo, Zehua; Liu, Sen; Tam, Adrian S.-W.; Xi, Kang; Chao, H. Jonathan

doi:10.1109/jsac.2019.2927072

Cited by 11 publications

(6 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to the expansion of the road traffic system, the continuous increase in bandwidth, processing capacity, and other technical innovations has expanded the physical limits of volume and speed in data transmission: from fiber‐optic communication first installed in Germany in 1993 and soon called “Datenautobahn” (data highway) to current 5G and future 6G wireless technologies. However, even the latest technological innovations face problems of physical limitations, for example, when “the limited buffer space on the commodity switches becomes the critical resource” (Xu et al., 2019, p. 1911), and they still suffer from latencies and congestions, for example, due to physical blockage and misalignment of signals (Poorzare & Augé, 2020, p. 176395). While the analogical reasoning should not be overused, there are obvious similarities here both between capacity increases on “highways” for vehicles and for data packets and between their impairment by limited traffic flow on feeder roads or limited data packet processing at hardware nodes, such as commodity switches.…”

Section: Historical and Relational Analogy: Massification And Congest...mentioning

confidence: 99%

Smart cities at risk: Systemic risk drivers in the blind spot of long‐term governance

Ottenburger

Ufer

2023

Risk Analysis

View full text Add to dashboard Cite

In this article, we analyze “digital massification” in smart cities, that is, an ever‐growing number of market participants, consumers, and Internet of Things devices with simultaneous accommodation of users to increasing disturbances and inconveniences due to data congestion—as a driver for systemic risk. We argue that digital massification phenomena largely escape societal awareness due to their protracted evolution and are therefore still in the blind spot of long‐term governance. Our analysis makes methodological use of historical and relational analogy, and we introduce and elaborate concepts and terms that allow us to discuss the evolutionary nature of massification, that is, the foreseeable increasing probability of the occurrence of trigger events. Using the analogy to the history of road traffic congestion, we deduce that digital massification will most likely lead to a future “risk transition” where tolerated disturbances and inconveniences of the present will turn into systemic impacts. This insight calls for heightened sensitivity in governance to massification phenomena to ensure the long‐term resilience of smart cities.

show abstract

Section: Historical and Relational Analogy: Massification And Congest...mentioning

confidence: 99%

Smart cities at risk: Systemic risk drivers in the blind spot of long‐term governance

Ottenburger

Ufer

2023

Risk Analysis

View full text Add to dashboard Cite

show abstract

“…RTT in traditional TCP/IP networks differs by several orders of magnitude from the block transfer time in data center networks, which means that RTT in traditional IP networks is no longer suitable for data center networks with high latency requirements. To this end, RAPID [28], T-Racks [29], and TCP-EFR [30] have focused on improving the fast retransmission/fast recovery mechanism in traditional networks, starting with repeated ACK, to avoid triggering RTO in the event of congestion and optimize short-flow FCT in response to congestion.…”

Section: Related Workmentioning

confidence: 99%

PFDCT: An Enhanced Transport Layer Protocol for Precise Flow Control in Data Centers

Wang

et al. 2023

Electronics

View full text Add to dashboard Cite

This paper proposes an enhanced transport layer protocol for precise flow control in data centers, referred to as PFDCT, combining explicit rate control (priority scheduling) and implicit rate control (marking algorithm with dynamic threshold) innovatively, while satisfying the transmission requirements of each data flow. It enables multiple data flows to share the bandwidth fairly, which is rarely achieved by previous data transmission protocols. Our design stems from the observation that different packets in a data center network need to be processed differentially and that the accuracy of the feedback of the indication information to the link state will directly affect the effectiveness of the congestion control mechanism. Our tests show that PFDCT has lower network jitter, better fairness, and significant performance gains over existing proposals (DCTCP, ICTCP, and L2DCT) in terms of Flow Completion Time (FCT) and throughput on average. Furthermore, we discussed the impact of the marking algorithm with dynamic thresholds on the performance of PFDCT, and the results of the experiments show that the addition of the marking algorithm will further improve the performance of PFDCT in terms of packet loss rate, latency, and FCT.

show abstract

“…And more there is no single solution for all scenarios. With the existence of many solutions either transport-based, application-based, or SDN-based ( [4], [8], [16], [17]) to handle incast, relying only on analytical performance modeling is not a practical long-term solution. With its capability of not relying on any domain-specific assumptions, machine learning can then be leveraged to construct a generalized model via a uniform training method.…”

Section: B Motivationsmentioning

confidence: 99%

“…This prevents buffer overflow and subsequent timeouts. The work in [8] proposes an intelligent selective packet discarding at the switch level. This intelligent discarding ensures that the sender responds to packet loss, using fast retransmission/fast recovery instead of RTO, and then avoiding RTO's penalty.…”

Section: Introductionmentioning

confidence: 99%

Learning-based Incast Performance Inference in Software-Defined Data Centers

Nougnanke

Labit

Bruyère

et al. 2021

2021 24th Conference on Innovation in Clouds, Internet and Networks and Workshops (ICIN)

View full text Add to dashboard Cite

Incast traffic is a many-to-one communication pattern used in many applications, including distributed storage, web-search with partition/aggregation design pattern, and MapReduce, commonly in data centers. It is generally composed of short-lived flows that may be queued behind large flows' packets in congested switches where performance degradation is observed. Smart buffering at the switch level is sensed to mitigate this issue by automatically and dynamically adapting to traffic conditions changes in the highly dynamic data center environment. But for this dynamic and smart buffer management to become effectively beneficial for all the traffic, and especially for incast the most critical one, incast performance models that provide insights on how various factors affect it are needed. The literature lacks these types of models. The existing ones are analytical models, which are either tightly coupled with a particular protocol version or specific to certain empirical data. Motivated by this observation, we propose a machine-learning-based incast performance inference. With this prediction capability, smart buffering scheme or other QoS optimization algorithms could anticipate and efficiently optimize system parameters adjustment to achieve optimal performance. Since applying machine learning to networks managed in a distributed fashion is hard, the prediction mechanism will be deployed on an SDN control plane. We could then take advantage of SDN's centralized global view, its telemetry capabilities, and its management flexibility.

show abstract

RAPID: Avoiding TCP Incast Throughput Collapse in Public Clouds With Intelligent Packet Discarding

Cited by 11 publications

References 22 publications

Smart cities at risk: Systemic risk drivers in the blind spot of long‐term governance

Smart cities at risk: Systemic risk drivers in the blind spot of long‐term governance

PFDCT: An Enhanced Transport Layer Protocol for Precise Flow Control in Data Centers

Learning-based Incast Performance Inference in Software-Defined Data Centers

Contact Info

Product

Resources

About