TY - GEN
T1 - Landing Reinforcement Learning onto Smart Scanning of The Internet of Things
AU - Qu, Jian
AU - Ma, Xiaobo
AU - Liu, Wenmao
AU - Sang, Hongqing
AU - Li, Jianfeng
AU - Xue, Lei
AU - Luo, Xiapu
AU - Li, Zhenhua
AU - LI Feng,
AU - Guan, Xiaohong
N1 - Funding Information:
This work was supported in part by National Natural Science Foundation (61972313, 62002306), Postdoctoral Science Foundation (2019M663725, 2021T140543), Hong Kong RGC Project (No. PolyU15223918), CCF-NSFOCUS KunPeng Research Fund, Research Fund from Huawei Technology, PolyU startup fund (ZVU7), and the Fundamental Research Funds for the Central Universities, of China. Xiaobo Ma is an XJTU Tang Scholar supported by Cyrus Tang Foundation.
Publisher Copyright:
© 2022 IEEE.
PY - 2022/6/20
Y1 - 2022/6/20
N2 - Cyber search engines, such as Shodan and Censys, have gained popularity due to their strong capability of indexing the Internet of Things (IoT). They actively scan and fingerprint IoT devices for unearthing IP-device mapping. Because of the large address space of the Internet and the mapping’s mutative nature, efficiently tracking the evolution of IP-device mapping with a limited budget of scans is essential for building timely cyber search engines. An intuitive solution is to use reinforcement learning to schedule more scans to networks with high churn rates of IP-device mapping. However, such an intuitive solution has never been systematically studied. In this paper, we take the first step toward demystifying this problem based on our experiences in maintaining a global IoT scanning platform. Inspired by the measurement study of large-scale real-world IoT scan records, we land reinforcement learning onto a system capable of smartly scanning IoT devices in a principled way. We disclose key parameters affecting the effectiveness of different scanning strategies, and find that our system would achieve growing advantages with the proliferation of IoT devices.
AB - Cyber search engines, such as Shodan and Censys, have gained popularity due to their strong capability of indexing the Internet of Things (IoT). They actively scan and fingerprint IoT devices for unearthing IP-device mapping. Because of the large address space of the Internet and the mapping’s mutative nature, efficiently tracking the evolution of IP-device mapping with a limited budget of scans is essential for building timely cyber search engines. An intuitive solution is to use reinforcement learning to schedule more scans to networks with high churn rates of IP-device mapping. However, such an intuitive solution has never been systematically studied. In this paper, we take the first step toward demystifying this problem based on our experiences in maintaining a global IoT scanning platform. Inspired by the measurement study of large-scale real-world IoT scan records, we land reinforcement learning onto a system capable of smartly scanning IoT devices in a principled way. We disclose key parameters affecting the effectiveness of different scanning strategies, and find that our system would achieve growing advantages with the proliferation of IoT devices.
UR - http://www.scopus.com/inward/record.url?scp=85133267241&partnerID=8YFLogxK
U2 - 10.1109/INFOCOM48880.2022.9796737
DO - 10.1109/INFOCOM48880.2022.9796737
M3 - Conference article published in proceeding or book
T3 - Proceedings - IEEE INFOCOM
SP - 2088
EP - 2097
BT - INFOCOM 2022 - IEEE Conference on Computer Communications
PB - IEEE
ER -