Yu-Kuen Lai

Prof. Yu-Kuen Lai

orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 July 2019 in IEEE Access
Verified
Journal Article
Open access
TWTheophilus Wellem
Theophilus Wellem
Yu-Kuen LaiYu-Kuen Lai
Yu-Kuen Lai
CHChao-Yuan Huang
Chao-Yuan Huang
WCWen-Yaw Chung
Wen-Yaw Chung
Sketch-based data streaming algorithms are used in many network traffic monitoring applications to obtain accurate estimates of traffic flow statistics. However, the current implementation of sketch-based methods in network monitoring is too application-specific. Their flexibility is limited as the hardware implementation of sketch data structure in the network device depends on the measurement tasks. The sketch counters only summarize flow statistics that are relevant to a specific measurement task and cannot be reused for different measurement tasks. In this paper, we propose FlexSketchMon, a system designed to provide the flexibility of using various sketch-based algorithms for traffic monitoring and measurement tasks. FlexSketchMon leverages on a novel data plane architecture that collects traffic flow statistics and provides arbitrary flow aggregations to the monitoring applications. The data plane design comprises a flow counter table and a flow key table for storing flow-level data. FlexSketchMon is implemented on the NetFPGA-SUME platform and is capable of processing network traffic at line rate in a worst-case scenario corresponding to a 64-byte minimum Ethernet frame size. The update of the flow counter table, which is the critical path in the proposed system, can achieve a throughput of 96 Gbps. The simulation results based on a real-world network traffic traces for three monitoring applications-estimation, superspreader detection, and heavy hitter detection-are presented to demonstrate the performance of FlexSketchMon. The results show that FlexSketchMon yields comparable and better measurement accuracy compared to previous approaches.

1 July 2025 in Electronics
Verified
Journal Article
Open access
PDF file
DHDrixter Velayo Hernandez
Drixter Velayo Hernandez
Yu-Kuen LaiYu-Kuen Lai
Yu-Kuen Lai
HIHargyo T. N. Ignatius
Hargyo T. N. Ignatius
The exponential growth of Distributed Denial-of-Service (DDoS) attacks in high-speed networks presents significant real-time detection and mitigation challenges. The existing detection frameworks are categorized into flow-based and packet-based detection approaches. Flow-based approaches usually suffer from high latency and controller overhead in high-volume traffic. In contrast, packet-based approaches are prone to high false-positive rates and limited attack classification, resulting in delayed mitigation responses. To address these limitations, we propose a real-time DDoS detection architecture that combines hardware-accelerated statistical preprocessing with GPU-accelerated deep learning models. The raw packet header information is transformed into multivariate time series data to enable classification of complex traffic patterns using Temporal Convolutional Networks (TCN), Long Short-Term Memory (LSTM) networks, and Transformer architectures. We evaluated the proposed system using experiments conducted under low to high-volume background traffic to validate each model’s robustness and adaptability in a real-time network environment. The experiments are conducted across different time window lengths to determine the trade-offs between detection accuracy and latency. The results show that larger observation windows improve detection accuracy using TCN and LSTM models and consistently outperform the Transformer in high-volume scenarios. Regarding model latency, TCN and Transformer exhibit constant latency across all window sizes. We also used SHAP (Shapley Additive exPlanations) analysis to identify the most discriminative traffic features, enhancing model interpretability and supporting feature selection for computational efficiency. Among the experimented models, TCN achieves the most balance between detection performance and latency, making it an applicable model for the proposed architecture. These findings validate the feasibility of the proposed architecture and support its potential as a real-time DDoS detection application in a realistic high-speed network.

27 September 2022 in IEEE Access
Verified
Journal Article
Open access
Yu-Kuen LaiYu-Kuen Lai
Yu-Kuen Lai
CTCheng-Lin Tsai
Cheng-Lin Tsai
CCCheng-Han Chuang
Cheng-Han Chuang
XKXiu-Wen Ku
Xiu-Wen Ku
JCJim Hao Chen
Jim Hao Chen
The empirical entropy of the network flow attributes is an essential measure for identifying anomalous network traffic. However, computing the exact entropy values for high-speed networks in real-time is computationally expensive. Accordingly, the present study replaces the complex computations of existing stable random projection methods for entropy estimation with a simple table lookup procedure. Notably, the size of the lookup table is reduced through a piece-wise linear interpolation heuristic in order to facilitate the implementation of the proposed scheme in resource-constrained pipeline environments. The proposed architecture enables entropy estimation to be performed using both the Log-Mean Estimator (LME) method and the New Estimator of Compressed Counting (NECC) algorithm reported in the literature. The feasibility of the proposed approach is verified empirically using both real-world network traffic traces and synthetic data streams. Moreover, the practical applicability is demonstrated via stream-based implementation in the programmable data planes of the NetFPGA-Plus framework and a Tofino P4 switch, respectively. The results indicate that the proposed tabulation-based entropy estimation scheme allows minimum-sized Ethernet frames to be processed with a wire speed of up to several hundred gigabits per second.

27 June 2019 in IEEE Systems Journal
Verified
Journal Article
Closed access
YLYing-Dar Lin
Ying-Dar Lin
Yu-Kuen LaiYu-Kuen Lai
Yu-Kuen Lai
YTYung-Liang Tsou
Yung-Liang Tsou
YLYuan-Cheng Lai
Yuan-Cheng Lai
ELEn-Cheng Liou
En-Cheng Liou
YCYita Chiang
Yita Chiang
Programmable control plane in software-defined networking (SDN), plays an essential role in the SDN architecture. The network function provided by the specialized hardware in a legacy network can be created in the form of software-based “SDN application” running on the controllers to manipulate entire network configurations. Therefore, the risk of having software bugs and errors in the SDN applications may threaten the normal operations of SDN networks. This paper presents systematic validation criteria and test cases based on the proposed novel methodologies for SDN application testing. The test framework can perform testbed build-up, generate desired packet sequences, and analyze results automatically. According to the results of a generic test suite, several issues are unveiled in the application under test (AUT). Some AUTs, which need to check all the incoming packets from OpenFlow switches, fail to meet the test criteria of burst packet-in and flow self-recycling. For most of the applications based on the Ryu controller, the evaluation results reveal that some are unable to recycle flow entries after they are unloaded. It is recommended that all flows populated by SDN applications must have timeout value specified to prevent unnecessary entries kept in the flow table.