Modern data centers are hosting a great number of various applications (e.g. MapReduce and web search) that require a high fan-in data communication, which easily causes serious packet losses and timeouts, substantially degrading the application performance. To address this issue, various host-based and switch-based transport protocols are proposed to eliminate timeout and improve the user experience. Unfortunately, although existing transport protocols can effectively eliminate the timeout, they inevitably result in persistent queueing backlog and degrade the network performance, especially delay-sensitive short flows. To this end, we propose a general scheme with ultra-low latency called UL2to address the above problem. Concretely, the sender periodically estimates the queueing delay of each packet on the transmission path and senses the degree of congestion based on its measured result. Then the sender timely yet cautiously executes a pausing transmission operation based on measured queueing delay, guaranteeing fast elimination of queue delays and high link utilization. Our evaluation indicates that UL2can effectively eliminate queue backlog and reduce the queueing delay by more than 90%. Moreover, UL2enhances the performance of state-of-the-art transport protocols in terms of flow completion times by up to 44.98%.

Increasing diversity of applications and services are being migrated to modern data center networks (DCNs), and these applications and services are typically generating various combinations of long and short flows with or without deadlines. However, most existing flow scheduling solutions for DCNs either adopt single-queue strategy (e.g., D2TCP) that inevitably results in non-urgent flows blocking urgent flows or multi-queue mechanism (e.g., Aemon) that is at the cost of packet reordering. In this paper, we present DPOT, a Dynamic Priority-based Ordered Transmission mechanism that is aimed at minimizing the flow completion time and deadline missing rate. To avoid the urgent flows being blocked by the non-urgent flows, the DPOT switch adopts different priority queues to buffer packets for different types of flows. What’s more, when the switch detects that a data flow promotes the priority of its packets, it utilizes a disorder-free flow scheduling mechanism based on dynamic prioritization to make sure that packets of the same flow arrive at the receiver without reordering. Through a series of experimental tests, we demonstrate that DPOT can decrease the deadline miss rate by up to 95% while reducing flow completion time by up to 45% in comparison to the state-of-the-art flow scheduling approaches.

Learning-based Multi-view Stereo (MVS) and stereo matching methods typically construct 3D cost volumes based on the camera frustum of the reference view. Regularization and regression of the cost volume are performed to obtain a depth map. However, the resolution of the output depth map is limited by the computational cost, and when performing feature extraction, the characteristics of convolution local perception make it impossible to capture global context information. In this paper, we propose CTPMVSNet by using the Global Feature Aware Transformer (GFT) to aggregate global context information within and across images. In order to make better use of GFT, we use Deformable Convolution Module (DCM) to ensure a smooth transition of the extracted feature range. In addition, in the cost volume regularization stage, to improve efficiency and generation accuracy, we design a lightweight regularization network with integrated pseudo-three-dimensional convolution, and our experiments on multiple dataset have achieved promising results.