Yingying Su (Tsinghua university), Dan Li (Tsinghua university), Li Chen (Zhongguancun Laboratory), Qi Li (Tsinghua university), Sitong Ling (Tsinghua University)

Although Resource Public Key Infrastructure (RPKI) is critical for securing inter-domain routing, we find that its key component, the RPKI Repository, is under studied. We conduct the first data-driven analysis of the existing RPKI Repository infrastructure, including a survey of worldwide AS administrators and a large-scale measurement of the existing RPKI Repository. Based on the findings of this study, we identify three key problems. Firstly, misbehaving RPKI authorities can easily manipulate RPKI objects, and Internet Number Resources (INRs) holders and Relying Parties (RPs) can neither prevent malicious behaviors of misbehaving authorities nor hold them accountable. Secondly, RPKI Repository is sensitive to failures: An attack or downtime of any Publication Point (PP) will prevent RPs from obtaining complete RPKI object views. Finally, we identify scalability issues with the current RPKI Repository, which are expected to worsen with the further deployment of Route Origin Authorization (ROA).

To address these problems, we propose dRR, an architecture that enhances the security, robustness, and scalability of the RPKI Repository while being compatible with standard RPKI. By introducing two new entities: Certificate Servers (CSs) and Monitors, dRR forms a decentralized federation of CSs, which enables the RPKI Repository to proactively defend against malicious behavior from authorities and to tolerate PPs' failures. dRR is also scalable for future large-scale deployment. We present the design of dRR in detail and implement a prototype of dRR on a global Internet testbed spanning 15 countries. Experimental results show that, although new security features are introduced, dRR only incurs negligible latency for certificate issuance and revocation. The throughput of certificate updates achieved by dRR is 450 times higher than the current maximum RPKI certificate update frequency.

View More Papers

Overconfidence is a Dangerous Thing: Mitigating Membership Inference Attacks...

Zitao Chen (University of British Columbia), Karthik Pattabiraman (University of British Columbia)

Read More

Aligning Confidential Computing with Cloud-native ML Platforms

Angelo Ruocco, Chris Porter, Claudio Carvalho, Daniele Buono, Derren Dunn, Hubertus Franke, James Bottomley, Marcio Silva, Mengmei Ye, Niteesh Dubey, Tobin Feldman-Fitzthum (IBM Research)

Read More

LMSanitator: Defending Prompt-Tuning Against Task-Agnostic Backdoors

Chengkun Wei (Zhejiang University), Wenlong Meng (Zhejiang University), Zhikun Zhang (CISPA Helmholtz Center for Information Security and Stanford University), Min Chen (CISPA Helmholtz Center for Information Security), Minghu Zhao (Zhejiang University), Wenjing Fang (Ant Group), Lei Wang (Ant Group), Zihui Zhang (Zhejiang University), Wenzhi Chen (Zhejiang University)

Read More

“I used to live in Florida”: Exploring the Impact...

Imani N. S. Munyaka (University of California, San Diego), Daniel A Delgado, Juan Gilbert, Jaime Ruiz, Patrick Traynor (University of Florida)

Read More