Rethinking Fast and Friendly Transport in Data Center Networks

Zhang, Tao; Huang, Jiawei; Chen, Kai; Wang, Jianxin; Chen, Jianer; Pan, Yi; Min, Geyong

doi:10.1109/tnet.2020.3012556

Cited by 27 publications

(10 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The tail latency for short RPCs (remote procedure calls) has been evaluated with a load close to 100%. Similarly, in [ 22 ], the performance of data center transport protocol has been investigated with a two-dimensional explicit congestion notification along with an analytical model to assess the convergence process.…”

Section: Related Workmentioning

confidence: 99%

Delay Analysis in IoT Sensor Networks

Althoubi

Alshahrani

Peyravi

2021

Sensors

View full text Add to dashboard Cite

Internet of Things (IoT) devices, particularly those used for sensor networks, are often latency-sensitive devices. The topology of the sensor network largely depends on the overall system application. Various configurations include linear, star, hierarchical and mesh in 2D or 3D deployments. Other applications include underwater communication with high attenuation of radio waves, disaster relief networks, rural networking, environmental monitoring networks, and vehicular networks. These networks all share the same characteristics, including link latency, latency variation (jitter), and tail latency. Achieving a predictable performance is critical for many interactive and latency-sensitive applications. In this paper, a two-stage tandem queuing model is developed to estimate the average end-to-end latency and predict the latency variation in closed forms. This model also provides a feedback mechanism to investigate other major performance metrics, such as utilization, and the optimal number of computing units needed in a single cluster. The model is applied for two classes of networks, namely, Edge Sensor Networks (ESNs) and Data Center Networks (DCNs). While the proposed model is theoretically derived from a queuing-based model, the simulation results of various network topologies and under different traffic conditions prove the accuracy of our model.

show abstract

Section: Related Workmentioning

confidence: 99%

Delay Analysis in IoT Sensor Networks

Althoubi

Alshahrani

Peyravi

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Since ECN-based schemes fail to perceive the transient congestion and the delay-based schemes is oversensitive to congestion, making them less competitive when coexisting with the traditional TCP, FFC [75] combines these twodimensional signals to achieve both fast convergence and TCP friendliness. The experimental results show that FFC converges fast and reduces the flow completion time significantly compared with DX and DCTCP.…”

Section: A Deadline-agnostic Schemesmentioning

confidence: 99%

“…Mix-based schemes combine the ECN and delay as the congestion signal. ECN signal effectively prevents packet loss, while delay signal can better control end-to-end queuing latency [75]. Thus, mix-based schemes can meet the demand of low flow completion time and high network throughput.…”

Section: F Summarymentioning

confidence: 99%

Survey on Traffic Management in Data Center Network: From Link Layer to Application Layer

Liu

Wang

et al. 2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

Due to the explosive growth of all kinds of Internet services, data centers have become an irreplaceable and vital infrastructure to support this soaring trend. Compared with traditional networks, data center networks (DCNs) have unique features, such as high bandwidth, low latency, many-to-one communication mode, shallow buffered switches, and multi-root topology. These new characteristics pose a lot of challenges to previous network technics (e.g., Ethernet, Equal Cost Multi-path (ECMP), TCP), making them hard to adapt to DCNs and leading to severe performance degradation. In order to solve these challenges, DCNs have attracted a lot of attention from the industry and academia in recent years, and many new mechanisms in different layers are proposed to improve the transmission performance of data center networks. In the meantime, many surveys have emerged currently to introduce the current research of data center networks. However, previous surveys of DCNs mainly focus on only one specific network layer, making them difficult for readers to know about advanced researches on a holistic level. To help readers comprehend the current research progress of data center networks quickly, we employ a multi-layered top down taxonomy to classify the literature and propose several probable dimensions for future research in this area.

show abstract

“…In recent years, researchers have designed plenty of transport protocols to improve the transmission performance of datacenter networks [1,2,3,4,5,6,18,22,23,24,25,26,27,28,29]. Nonetheless, with the sharp increase of network capacity, various load balancing mechanisms have also been proposed to facilitate parallel data transmission across multiple paths, thus further obtaining performance enhancements.…”

Section: Related Workmentioning

confidence: 99%

“…Guaranteeing application performance is crucial for providing good user experience in datacenters. Tons of studies have reported that optimizing the transmission performance of datacenter network (DCN) is the key [1,2,3,4,5,6]. Therefore, to boost the network capacity thus speeding up data transfer, modern DCNs are usually organized in multi-rooted tree topologies with rich parallel paths, such as leaf-spine [7,8,9,10,11], and split the application traffic among multiple available paths.…”

Section: Introductionmentioning

confidence: 99%

Fine-grained Load Balancing with Traffic-aware Rerouting in DataCenter Networks

Zhang

Lei

Zhang

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Modern datacenters provide a wide variety of application services, which generate a mix of delay-sensitive short flows and throughput-oriented long flows, transmitting in the multi-path datacenter network. Though the existing load balancing designs successfully make full use of available parallel paths and attain high bisection network bandwidth, they reroute flows regardless of their dissimilar performance requirements. The short flows suffer from the problems of large queuing delay and packet reordering, while the long flows fail to obtain high throughput due to low link utilization and packet reordering. To address these inefficiency, we design a fine-grained load balancing scheme, namely TR (Traffic-aware Rerouting), which identifies flow types and executes flexible and traffic-aware rerouting to balance the performances of both short and long flows. Besides, to avoid packet reordering, TR leverages the reverse ACKs to estimate the switch-to-switch delay, thus excluding paths that potentially cause packet reordering. Moreover, TR is only deployed on the switch without any modification on end-hosts. The experimental results of large-scale NS2 simulations show that TR reduces the average and tail flow completion time for short flows by up to 60% and 80%, as well as provides up to 3.02x gain in throughput of long flows compared to the state-of-the-art load balancing schemes.

show abstract

Rethinking Fast and Friendly Transport in Data Center Networks

Cited by 27 publications

References 33 publications

Delay Analysis in IoT Sensor Networks

Delay Analysis in IoT Sensor Networks

Survey on Traffic Management in Data Center Network: From Link Layer to Application Layer

Fine-grained Load Balancing with Traffic-aware Rerouting in DataCenter Networks

Contact Info

Product

Resources

About