Shihan Lin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Shihan Lin has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Xiaowei Yang
Duke University, Durham, USA
6 shared publications - Ramesh Govindan
University of Southern California, Los Angeles, California, USA
4 shared publications - Todd Arnold
U.S. Military Academy, West Point, NY, USA
4 shared publications - Xiao Zhang
Cisco ThousandEyes, Raleigh, NC, USA
4 shared publications - Yi Zhou
Duke University, Durham, NC, USA
4 shared publications - Yunming Xiao
The Chinese University of Hong Kong, Shenzhen, Shenzhen, China
2 shared publications
Latest Publications
- Open Access
- 10 April 2027
- Published by Association for Computing Machinery (ACM)
- p. 692-707
PhantomBox: Fully Homomorphic Web Application Firewall
Web Application Firewalls (WAFs) defend against malicious traffic by inspecting application payloads against regular expression (RegEx) rulesets. Because ruleset maintenance requires specialized expertise, operators typically delegate WAF functionality to third-party middleboxes, which cannot inspect encrypted traffic without access to the secret keys. Prior privacy-preserving WAFs sacrifice either RegEx expressiveness, third-party offloading, or communication costs. Fully Homomorphic Encryption (FHE) can evaluate RegEx on encrypted traffic, but thousand-fold ciphertext expansion and seconds-per-character compute costs have kept FHE-based WAFs unrealistic. We present PhantomBox, an FHE-based WAF that co-designs the datapath from client-side encryption through middlebox rule compilation and execution. To mitigate the compute bottleneck, PhantomBox introduces a novel ahead-of-time RegEx compiler with two FHE-specific optimizations: Fast Literal Matching and Boolean Fusion. The compiler also exploits equality saturation to further leverage the optimizations. To reduce network overhead, PhantomBox employs hardware-accelerated transciphering, keeping on-wire size comparable to plaintext while reducing client encryption latency by 84.6× over CPU. Evaluated on the Proofpoint ruleset with over 5,800 web-attack signatures, PhantomBox achieves a median evaluation latency of 7.7 ms on an AWS EC2 F2 instance. PhantomBox is the first system to make FHE-based WAF practical for real-world deployment, achieving full RegEx expressiveness, middlebox offloading, and cryptographic privacy simultaneously.
- 10 April 2027
- Published by Association for Computing Machinery (ACM)
- p. 692-707
Open AccessPhantomBox: Fully Homomorphic Web Application Firewall
Web Application Firewalls (WAFs) defend against malicious traffic by inspecting application payloads against regular expression (RegEx) rulesets. Because ruleset maintenance requires specialized expertise, operators typically delegate WAF functionality to third-party middleboxes, which cannot inspect encrypted traffic without access to the secret keys. Prior privacy-preserving WAFs sacrifice either RegEx expressiveness, third-party offloading, or communication costs. Fully Homomorphic Encryption (FHE) can evaluate RegEx on encrypted traffic, but thousand-fold ciphertext expansion and seconds-per-character compute costs have kept FHE-based WAFs unrealistic. We present PhantomBox, an FHE-based WAF that co-designs the datapath from client-side encryption through middlebox rule compilation and execution. To mitigate the compute bottleneck, PhantomBox introduces a novel ahead-of-time RegEx compiler with two FHE-specific optimizations: Fast Literal Matching and Boolean Fusion. The compiler also exploits equality saturation to further leverage the optimizations. To reduce network overhead, PhantomBox employs hardware-accelerated transciphering, keeping on-wire size comparable to plaintext while reducing client encryption latency by 84.6× over CPU. Evaluated on the Proofpoint ruleset with over 5,800 web-attack signatures, PhantomBox achieves a median evaluation latency of 7.7 ms on an AWS EC2 F2 instance. PhantomBox is the first system to make FHE-based WAF practical for real-world deployment, achieving full RegEx expressiveness, middlebox offloading, and cryptographic privacy simultaneously.
- Open Access
- 19 November 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Networking
- Vol. 4 (CoNEXT4) , 1-24
Latency-Aware AS Path Prepending for Cloud Routing
Despite the traffic engineering efforts of cloud providers to reduce latency for cloud services, both researchers and practitioners observe that circuitous routing persists and that substantial opportunities for improvement remain. Lowering latency through traffic engineering is difficult because the Internet's inter-domain routing protocol, the Border Gateway Protocol (BGP), is oblivious to any performance metric. Moreover, changing BGP itself has historically been a slow and arduous process. In this work, we investigate how a cloud provider can introduce a latency metric into BGP with minimal changes and minimal disruption to the existing routing paradigm. We propose repurposing an existing attribute, the AS path, to encode latency through latency-proportional AS-path prepending (ASPP), requiring no changes to the protocol itself. To make this method practical, we further propose an AS selection algorithm that helps a cloud provider identify a small set of direct and near-direct neighbors at which to deploy the method. To evaluate our proposal, we carefully develop a simulation-based methodology that fairly compares ASPP against the baseline. Our evaluation shows that ASPP is well suited for partial deployment: it approximates the performance gains of a global deployment with only 70 participating ASes, and improves 19.4% to 32.5% more client paths than the previously proposed BGP attribute extension, AIGP. We argue that, given these performance gains, cloud providers are natural first adopters of the proposed mechanism, as they reap immediate benefits and need only negotiate deployment with a small set of ASes.
- 19 November 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Networking
- Vol. 4 (CoNEXT4) , 1-24
Open AccessLatency-Aware AS Path Prepending for Cloud Routing
Despite the traffic engineering efforts of cloud providers to reduce latency for cloud services, both researchers and practitioners observe that circuitous routing persists and that substantial opportunities for improvement remain. Lowering latency through traffic engineering is difficult because the Internet's inter-domain routing protocol, the Border Gateway Protocol (BGP), is oblivious to any performance metric. Moreover, changing BGP itself has historically been a slow and arduous process. In this work, we investigate how a cloud provider can introduce a latency metric into BGP with minimal changes and minimal disruption to the existing routing paradigm. We propose repurposing an existing attribute, the AS path, to encode latency through latency-proportional AS-path prepending (ASPP), requiring no changes to the protocol itself. To make this method practical, we further propose an AS selection algorithm that helps a cloud provider identify a small set of direct and near-direct neighbors at which to deploy the method. To evaluate our proposal, we carefully develop a simulation-based methodology that fairly compares ASPP against the baseline. Our evaluation shows that ASPP is well suited for partial deployment: it approximates the performance gains of a global deployment with only 70 participating ASes, and improves 19.4% to 32.5% more client paths than the previously proposed BGP attribute extension, AIGP. We argue that, given these performance gains, cloud providers are natural first adopters of the proposed mechanism, as they reap immediate benefits and need only negotiate deployment with a small set of ASes.
- Open Access
- 11 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 447-460
XFir: Accelerating New-Flow Setup on Host Servers of a Large Cloud Network
- 11 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 447-460
Open AccessXFir: Accelerating New-Flow Setup on Host Servers of a Large Cloud Network
- Open Access
- 16 June 2025
- Published by Association for Computing Machinery (ACM) in ACM SIGMETRICS Performance Evaluation Review
- Vol. 53 (1) , 73-75
Tiered Cloud Routing: Methodology, Latency, and Improvement
Large cloud providers including AWS, Azure, and Google offer two tiers of network services to their customers: WAN-transit service and inet-transit service. Little is known about how each cloud provider offers different transit services, how well these services work, and whether the quality of those services can be further improved. In this work, we conduct a large-scale study to answer these questions. Using RIPE Atlas probes as vantage points, we explore how traffic enters and leaves the WAN of each of the three clouds. In addition, we measure the access latencies of these two network services of each cloud and compare them with emulated alternative routing strategies.
- 16 June 2025
- Published by Association for Computing Machinery (ACM) in ACM SIGMETRICS Performance Evaluation Review
- Vol. 53 (1) , 73-75
Open AccessTiered Cloud Routing: Methodology, Latency, and Improvement
Large cloud providers including AWS, Azure, and Google offer two tiers of network services to their customers: WAN-transit service and inet-transit service. Little is known about how each cloud provider offers different transit services, how well these services work, and whether the quality of those services can be further improved. In this work, we conduct a large-scale study to answer these questions. Using RIPE Atlas probes as vantage points, we explore how traffic enters and leaves the WAN of each of the three clouds. In addition, we measure the access latencies of these two network services of each cloud and compare them with emulated alternative routing strategies.
- Open Access
- 9 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 73-75
Tiered Cloud Routing: Methodology, Latency, and Improvement
- 9 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 73-75
Open AccessTiered Cloud Routing: Methodology, Latency, and Improvement
- Open Access
- 6 March 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Measurement and Analysis of Computing Systems
- Vol. 9 (1) , 1-41
Tiered Cloud Routing: Methodology, Latency, and Improvement
Large cloud providers including AWS, Azure, and Google Cloud offer two tiers of network services to their customers: one class uses the providers' private wide area networks (WAN-transit) to carry a customer's traffic as much as possible, and the other uses the public internet (inet-transit). Little is known about how each cloud provider configures its network to offer different transit services, how well these services work, and whether the quality of those services can be further improved. In this work, we conduct a large-scale study to answer these questions. Using RIPE Atlas probes as vantage points, we explore how traffic enters and leaves each cloud's WAN. In addition, we measure the access latency of the WAN-transit and the inet-transit service of each cloud and compare it with that of an emulated performance-based routing strategy. Our study shows that despite the cloud providers' intention to carry customers' traffic on its WAN to the maximum extent possible, for about 12% (Azure) and 13% (Google) of our vantage points, traffic exits the cloud WAN early at cloud edges more than 5000km away from the vantage points' nearest cloud edges. In contrast, more than 84% (AWS), 78% (Azure), and 81% (Google) of vantage points enter a cloud WAN within a 500km radius of their respective locations. Moreover, we find that cloud providers employ different routing strategies to implement the inet-transit service, leading to transit policies that may deviate from their advertised service descriptions. Finally, we find that a performance-based routing strategy can significantly reduce latencies in all three cloud providers for 4% to 85% of vantage point and cloud region pairs.
- 6 March 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Measurement and Analysis of Computing Systems
- Vol. 9 (1) , 1-41
Open AccessTiered Cloud Routing: Methodology, Latency, and Improvement
Large cloud providers including AWS, Azure, and Google Cloud offer two tiers of network services to their customers: one class uses the providers' private wide area networks (WAN-transit) to carry a customer's traffic as much as possible, and the other uses the public internet (inet-transit). Little is known about how each cloud provider configures its network to offer different transit services, how well these services work, and whether the quality of those services can be further improved. In this work, we conduct a large-scale study to answer these questions. Using RIPE Atlas probes as vantage points, we explore how traffic enters and leaves each cloud's WAN. In addition, we measure the access latency of the WAN-transit and the inet-transit service of each cloud and compare it with that of an emulated performance-based routing strategy. Our study shows that despite the cloud providers' intention to carry customers' traffic on its WAN to the maximum extent possible, for about 12% (Azure) and 13% (Google) of our vantage points, traffic exits the cloud WAN early at cloud edges more than 5000km away from the vantage points' nearest cloud edges. In contrast, more than 84% (AWS), 78% (Azure), and 81% (Google) of vantage points enter a cloud WAN within a 500km radius of their respective locations. Moreover, we find that cloud providers employ different routing strategies to implement the inet-transit service, leading to transit policies that may deviate from their advertised service descriptions. Finally, we find that a performance-based routing strategy can significantly reduce latencies in all three cloud providers for 4% to 85% of vantage point and cloud region pairs.
- Open Access
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 834-835
Demo: CTSim: A Scalable and Flexible Cybertwin Network Simulator for Internet of Things Scenarios
The future of networking must address the connectivity demands of billions of people and trillions of networked devices. The Intelligent Internet of Everything (IoE) is widely considered the next evolution of the Internet, yet it poses significant challenges to the current TCP/IP architecture. To overcome these challenges, the concept of the Cybertwin network has been proposed as a promising future Internet architecture. To facilitate the advancement of Cybertwin-related research, we have developed CTSim, a Cybertwin network simulator. Our demo demonstrates that the Cybertwin network could effectively enhances mobility, availability, and security in Internet of Things scenarios, making it a powerful tool for researchers in these fields.
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 834-835
Open AccessDemo: CTSim: A Scalable and Flexible Cybertwin Network Simulator for Internet of Things Scenarios
The future of networking must address the connectivity demands of billions of people and trillions of networked devices. The Intelligent Internet of Everything (IoE) is widely considered the next evolution of the Internet, yet it poses significant challenges to the current TCP/IP architecture. To overcome these challenges, the concept of the Cybertwin network has been proposed as a promising future Internet architecture. To facilitate the advancement of Cybertwin-related research, we have developed CTSim, a Cybertwin network simulator. Our demo demonstrates that the Cybertwin network could effectively enhances mobility, availability, and security in Internet of Things scenarios, making it a powerful tool for researchers in these fields.
- Open Access
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 130-138
Browsing without Third-Party Cookies: What Do You See?
Third-party web cookies are often used for privacy-invasive behavior tracking. Partly due to privacy concerns, browser vendors have started to block all third-party cookies in recent years. To understand the effects of such third-party cookieless browsing, we crawled and measured the top 10,000 Tranco websites. We developed a framework to remove third-party cookies and analyze the differences between the appearance of web pages with and without these cookies. We find that disabling third-party cookies has no substantial effect on website appearance including layouts, text, and images. This validates the industry-wide shift towards cookieless browsing as a way to protect user privacy without compromising on the user experience.
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 130-138
Open AccessBrowsing without Third-Party Cookies: What Do You See?
Third-party web cookies are often used for privacy-invasive behavior tracking. Partly due to privacy concerns, browser vendors have started to block all third-party cookies in recent years. To understand the effects of such third-party cookieless browsing, we crawled and measured the top 10,000 Tranco websites. We developed a framework to remove third-party cookies and analyze the differences between the appearance of web pages with and without these cookies. We find that disabling third-party cookies has no substantial effect on website appearance including layouts, text, and images. This validates the industry-wide shift towards cookieless browsing as a way to protect user privacy without compromising on the user experience.