Chi-Hsiang Hung
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Chi-Hsiang Hung has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Charles H. -P. Wen
National Yang Ming Chiao Tung University, Hsinchu, Taiwan, Taiwan
4 shared publications - Michael I.-C. Wang
Department of Electrical and Computer Engineering, New York University Tandon School of Engineering, Brooklyn, New York, NY, USA
2 shared publications Alan Y.-P. Lee
1 shared publicationChain-Wu Lee
Network Infrastructure Design Center, Inventec Corporation
1 shared publicationJheng-Jyun Wang
Department of Electrical and Computer Engineering, National Chiao Tung University
1 shared publication- Kuochen Wang
Department of Computer Science, National Chiao Tung University, Hsinchu 300, Taiwan
1 shared publication
Latest Publications
- Open Access
- 4 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 183089-183102
Mitigating Microbursts by Packet Recirculation in Programmable Switch
Microbursts pose a significant challenge in 5G networks, leading to packet drops and high latency that impact user experience. Existing mitigation systems often suffer from transmission delays, high packet drop rates, and potential failure due to improper threshold settings. To address these challenges, we propose Programmable Switch-based Microburst Mitigation (PSMM), an innovative solution that employs two key techniques: in-switch recirculation and analytical threshold determination. The in-switch recirculation technique enables buffering packets within the switch during microburst events, eliminating the need for packet detours to neighboring switches and thereby reducing transmission delay and packet drops. The analytical threshold determination technique provides an optimal threshold range to trigger mitigation, preventing mitigation failure. Our extensive experiments demonstrate that PSMM can significantly outperform existing systems, reducing packet drops by 97.8% and latency by 26%.The findings highlight the effectiveness of PSMM in utilizing packet recirculation to mitigate microbursts and improve network performance and reliability in 5G environments.
- 4 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 183089-183102
Open AccessMitigating Microbursts by Packet Recirculation in Programmable Switch
Microbursts pose a significant challenge in 5G networks, leading to packet drops and high latency that impact user experience. Existing mitigation systems often suffer from transmission delays, high packet drop rates, and potential failure due to improper threshold settings. To address these challenges, we propose Programmable Switch-based Microburst Mitigation (PSMM), an innovative solution that employs two key techniques: in-switch recirculation and analytical threshold determination. The in-switch recirculation technique enables buffering packets within the switch during microburst events, eliminating the need for packet detours to neighboring switches and thereby reducing transmission delay and packet drops. The analytical threshold determination technique provides an optimal threshold range to trigger mitigation, preventing mitigation failure. Our extensive experiments demonstrate that PSMM can significantly outperform existing systems, reducing packet drops by 97.8% and latency by 26%.The findings highlight the effectiveness of PSMM in utilizing packet recirculation to mitigate microbursts and improve network performance and reliability in 5G environments.
- 10 November 2023
- Published by Elsevier in Future Generation Computer Systems
- Vol. 152, 333-342
PS-IPS: Deploying Intrusion Prevention System with machine learning on programmable switch
- 10 November 2023
- Published by Elsevier in Future Generation Computer Systems
- Vol. 152, 333-342
PS-IPS: Deploying Intrusion Prevention System with machine learning on programmable switch
- 25 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
P4SF: A High-Performance Stateful Firewall on Commodity P4-Programmable Switch
This paper presents a high-performance stateful firewall called P4SF that runs on a commodity P4-programmable switch and uses an extended finite state machine to provide match-state-action in the forwarding plane for stateful processing while significantly reducing the controller’s workload. P4SF is composed of three key blocks (i.e. Match Block, State Block, Action Block) that are responsible for reading/writing flow states, maintaining state transitions, and forwarding packets. Preemptive data caching is also realized into a buffer called State Pre-Fetch in P4SF for hiding transmission delay during state updates of flows. As a result, P4SF is successfully exercised on a commodity P4-programmable switch, and can be scaled to support 384,000 entries (120,000 under the three-way handshake in TCP connections) for TCP flows, achieving the 100Gb/s linerate speed for packet forwarding.
- 25 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
P4SF: A High-Performance Stateful Firewall on Commodity P4-Programmable Switch
This paper presents a high-performance stateful firewall called P4SF that runs on a commodity P4-programmable switch and uses an extended finite state machine to provide match-state-action in the forwarding plane for stateful processing while significantly reducing the controller’s workload. P4SF is composed of three key blocks (i.e. Match Block, State Block, Action Block) that are responsible for reading/writing flow states, maintaining state transitions, and forwarding packets. Preemptive data caching is also realized into a buffer called State Pre-Fetch in P4SF for hiding transmission delay during state updates of flows. As a result, P4SF is successfully exercised on a commodity P4-programmable switch, and can be scaled to support 384,000 entries (120,000 under the three-way handshake in TCP connections) for TCP flows, achieving the 100Gb/s linerate speed for packet forwarding.
- Open Access
- 30 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 2446-2456
Real-Time In-Network Microburst Mitigation on Programmable Switch
Microbursts in the datacenter network (DCN) last for an extremely short time in switches and are difficult to discover from a coarse-grained perspective. Most prior works are dedicated to in-network detection of microbursts and have not yet attempted to mitigate them in real time. Therefore, a realtime microburst mitigator, called RIMM, is proposed in this work and mainly applies orderly detours with flowlet intervals for preventing packet loss and packet retransmission in the network. RIMM mainly consists of three key components: (1) detour launcher, (2) packet sequencer, and (3) post-detour handler, and can entirely work in-network on a programmable switch. Experimental results demonstrate that RIMM is capable of preventing packet loss, allowing the reduction of packet retransmissions, while effectively reducing microbursts, resulting in an enhanced data-center network performance.
- 30 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 2446-2456
Open AccessReal-Time In-Network Microburst Mitigation on Programmable Switch
Microbursts in the datacenter network (DCN) last for an extremely short time in switches and are difficult to discover from a coarse-grained perspective. Most prior works are dedicated to in-network detection of microbursts and have not yet attempted to mitigate them in real time. Therefore, a realtime microburst mitigator, called RIMM, is proposed in this work and mainly applies orderly detours with flowlet intervals for preventing packet loss and packet retransmission in the network. RIMM mainly consists of three key components: (1) detour launcher, (2) packet sequencer, and (3) post-detour handler, and can entirely work in-network on a programmable switch. Experimental results demonstrate that RIMM is capable of preventing packet loss, allowing the reduction of packet retransmissions, while effectively reducing microbursts, resulting in an enhanced data-center network performance.
- 1 March 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 832-836
Heterogeneous Flow Table Integration for Capacity Enhancement in Software-Defined Networks
Commercial OpenFlow-enabled legacy switches implement flow tables with Ternary Content Addressable Memory (TCAM) to perform high speed flow matching process. However, the size of TCAM-based flow table is restricted to few thousands of entries due to manufacturing cost and high power consumption. In this paper, we propose the Flow Entry Agent which can integrate heterogeneous flow tables, such as ASIC-side MAC/IP/ACL tables and CPU-side SRAM-based tables to increase the storage capacity without modifying ASIC. Besides, the Agent will compress the CPU-side tables and perform flow matching process with these tables. Simulation results show that our Agent can not only increase the capacity of flow table to 11.38x to 65.24x but also reduce 80.99% to 87.36% of Packet_In messages with additional 1 MB CPU-side SRAM.
- 1 March 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 832-836
Heterogeneous Flow Table Integration for Capacity Enhancement in Software-Defined Networks
Commercial OpenFlow-enabled legacy switches implement flow tables with Ternary Content Addressable Memory (TCAM) to perform high speed flow matching process. However, the size of TCAM-based flow table is restricted to few thousands of entries due to manufacturing cost and high power consumption. In this paper, we propose the Flow Entry Agent which can integrate heterogeneous flow tables, such as ASIC-side MAC/IP/ACL tables and CPU-side SRAM-based tables to increase the storage capacity without modifying ASIC. Besides, the Agent will compress the CPU-side tables and perform flow matching process with these tables. Simulation results show that our Agent can not only increase the capacity of flow table to 11.38x to 65.24x but also reduce 80.99% to 87.36% of Packet_In messages with additional 1 MB CPU-side SRAM.