A5066560659- Software-Defined Networks and 5G
- Caching and Content Delivery
- Advanced Optical Network Technologies
- Cloud Computing and Resource Management
- Image and Video Quality Assessment
- Complex Network Analysis Techniques
- Advanced Photonic Communication Systems
- IoT and Edge/Fog Computing
- Internet Traffic Analysis and Secure E-voting
- Antibiotic Resistance in Bacteria
- Food Safety and Hygiene
- Fluid Dynamics and Thin Films
- Power Transformer Diagnostics and Insulation
- Mitochondrial Function and Pathology
- RNA Research and Splicing
- Educational Technology and Assessment
- Cancer Diagnosis and Treatment
- Islanding Detection in Power Systems
- Global trade, sustainability, and social impact
- Consumer Behavior in Brand Consumption and Identification
- Salmonella and Campylobacter epidemiology
- Power Line Inspection Robots
- Telecommunications and Broadcasting Technologies
- Genetic Neurodegenerative Diseases
- Digital Marketing and Social Media
Tianjin Agricultural University
2024
Beijing University of Posts and Telecommunications
2019-2024
Alibaba Group (China)
2024
Zhejiang University
2019-2023
Sanya University
2022-2023
The Ohio State University
2023
The Affiliated Yongchuan Hospital of Chongqing Medical University
2023
Chongqing Medical University
2023
State Key Laboratory of Networking and Switching Technology
2019-2022
North China Electric Power University
2020-2022
The cloud gateway is essential in the public as central hub of traffic. We show that horizontal scaling software gateways, once sustainable for years, no longer future-proof facing massive scale and rapid growth today's cloud. root cause stagnant performance CPU core, which prone to be overloaded by heavy hitters traffic goes far beyond Moore's law. To address this, we propose \emph{Sailfish}, a cloud-scale multi-tenant multi-service accelerated programmable switches. new challenge large...
Network reliability becomes increasingly important in modern data center networks (DCNs). The DCNs are expected to work sustainably under internal failures and assist network operators troubleshooting them rapidly. However, some will happen silently with packets discarded without producing any explicit notification before causing tremendous damage the network. To troubleshoot these "gray failures", this work, we present a rapid gray failure detection localization mechanism based on recently...
The In-band Network Telemetry (INT) enables hop-by-hop device-internal state exposure for reliably maintaining and troubleshooting data center networks. For achieving network-wide telemetry, orchestration on top of the INT primitive is further required. One straightforward solution to flood probe packets into network topology maximum measurement coverage, which, however, leads huge bandwidth overhead. A refined leverage SDN controller collect carry out centralized probing path planning,...
Prerequisite relations among concepts are crucial for educational applications. However, it is difficult to automatically extract domain-specific and learn the prerequisite them without labeled data.In this paper, we first high-quality phrases from a set of data, identify by graph based ranking method. Then, propose an iterative relation learning framework, called iPRL, which combines model recovery leverage both concept pair features dependencies materials. In experiments, evaluated our...
In this paper, a numerical model of bubble evolution is established by studying the microstructure pressboard and physical process generation. model, porous structure fibers equivalent to microtubules, gases in bubbles are classified into two categories, ideal gas state equation used describe state. Then pressure conditions at boundary evaporation rate water taken as solution solve quantities equation. The provides new way calculate initial temperature escape (ITBE). And inside obtained...
s (
The evaluation of the web‐browsing quality experience (QoE) is difficult to complete through traditional methods (e.g. deducing formulas or setting thresholds) due diversity websites and their contents. To evaluate QoE a general way, authors propose web architecture based on machine learning, consisting two parts: traffic classification sub‐system prediction sub‐system. When evaluating user experience, first classifies packets generated by visiting website into flowthrough some fields in...
Visibility is essential for operating and troubleshooting intricate networks. In-band Network Telemetry (INT) has been embedded in the latest merchant silicons to offer high-precision device traffic state visibility. INT actually an underlying technique each instance covers only one monitoring path. The network-wide measurement coverage therefore requires a high-level orchestration provision multiple paths. An optimal path planning expected produce minimum number of paths with overlapping...
In-band Network Telemetry (INT) enables fine-grained network monitoring to ease the management of large-scale networks, which, however, relies on real-time collection a huge amount telemetry data through southbound interface. For example, INT upload rate 28-pod FatTree topology reaches 3Tbps under probe frequency 100 times/s, which is rather unacceptable since controller-switch link bandwidth limited. To mitigate overhead, in this work, we propose INT-filter, novel measurement architecture...
In-band Network Telemetry (INT) enables hop-by-hop device-internal state exposure for reliably maintaining and troubleshooting networks. For achieving network-wide telemetry, high-level orchestration over the INT primitive is further needed. Existing solutions either incur significant bandwidth overhead or fail to promptly handle link failures. In this work, we propose INT-label, a lightweight telemetry architecture without introducing probe packets. INT-label periodically labels device...
We present Zoonet, a proactive virtual network telemetry system for multi-tenant clouds. The requirements are to (1) cover hyper-scale networks with millions of tenants and VMs top tenants; (2) handle frequent topology changes due tenants' configuration through flexible APIs; (3) adapt heterogeneous middleboxes along the probing paths; (4) achieve VM-to-VM without breaking tenant privacy; (5) differentiate physical problems. argue existing solutions fail satisfy our needs either incomplete...
At present, public clouds have served millions of tenants. To provide reliable services, cloud vendors need to perceive health status the network by building a telemetry system detect possible failures. While systems for physical networks been extensively studied, research on virtual is still insufficient. Different from networks, we conclude that faces new challenges feasibility, efficiency, and effectiveness. Specifically, 1) protect privacy tenants adapt heterogeneous middleboxes at data...
Ahead knowledge of link quality can reduce the energy consumption wireless sensor networks. In this paper, we propose a cloud reasoning-based prediction algorithm for A large number samples are collected from different scenarios, and their RSSI, LQI, SNR PRR parameters classified by self-adaptive Gaussian transformation algorithm. Taking limitation nodes? resources into consideration, Apriori is applied to determine association rules between physical layer parameters. reasoning that...
The lung is a common site for cancer metastasis. Some patients would develop metastases throughout the course of their illness. However, choosing surgical resection primary tumor (SRPT) or palliative treatment in with remains controversial.Lung metastatic diagnosed from 2010 to 2016 were selected Surveillance, Epidemiology, and End Results (SEER) database. Selected divided into two subgroups (surgery non-surgery). Further, all 58 types classified 13 subtypes. clinical demographic features...
In-band Network Telemetry (INT) enables hop-by-hop device-internal state exposure for maintaining and troubleshooting data center networks. To achieve <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">network-wide</i> telemetry coverage, orchestration on top of the INT primitive is required. A straightforward solution would flood network with probe packets maximum measurement which leads to a huge bandwidth overhead. refined leverages SDN...
The user perception of visiting a web page greatly affects the willingness to continue surfing website. However, it is difficult construct general Quality Experience (QoE) model via traditional methods (e.g. deducing formula or setting thresholds) because contents different websites are in endless varieties. To handle such an issue, we introduce WebQMon, QoE evaluation method using machine learning without any threshold, which easy deploy for websites. WebQMon can evaluate experience through...
Switchgear plays a vital role in the power system and is an essential part of ensuring safe stable operation system. In this paper, test platform for partial discharge switchgear built on laboratory site. The creeping model used insulators. At same time, ultrasonic signal collected when occurs switchgear. According to data obtained by platform, using two characteristic parameters shape scale, analyze discuss based threshold value, give value suggestion actual operating state. recommended...
Virtual private clouds provide sharing resources to a massive number of tenants for economics scale. In such clouds, off-the-shelf x86 boxes are widely deployed as network intermediate nodes. However, due rapid growth cloud traffic and significant slowdown CPU improvement in recent years, although horizontal scaling is still leveraged, overload packet losses caused by heavy hitters occasionally observed production environment, which seriously damage tenant's SLAs. To address this, we propose...
Deterministic latency is the key challenge that must be addressed in numerous 5G applications such as AR/VR. However, it difficult to make customized end-to-end resource reservation across multiple ISPs using IP-based QoS mechanisms. Information-Centric Networking (ICN) provides scalable and efficient content distribution at Internet scale due its in-network caching native multicast capabilities, deterministic can promisingly guaranteed by relevant objects appropriate locations. Existing...
In multi-tenant clouds, off-the-shelf x86 boxes are widely deployed as middleboxes. With the rapid growth of cloud traffic and migration to NFV deployment in recent years, CPU overloading at middleboxes becomes more an issue. From our data centers, we observed that was caused by heavy hitters. To address this issue, propose MIMIC, a cloud-scale flow backpressure system, implemented onto existing SmartNIC with FPGA acceleration. MIMIC rate-limits selected hitters through new per-flow protocol...