Unveiling IPv6 Scanning Dynamics: A Longitudinal Study Using Large Scale Proactive and Passive IPv6 Telescopes
Hammas Bin Tanveer, Echo Chan, Ricky K. P. Mok, Sebastian Kappes, Philipp Richter, Oliver Gasser, John Ronan, Arthur Berger, kc Claffy · 2025
Read Unveiling IPv6 Scanning Dynamics: A Longitudinal Study Using Large Scale Proactive and Passive IPv6 Telescopes by Hammas Bin Tanveer, Echo Chan, Ricky K. P. Mok, Sebastian Kappes, Philipp Richter, Oliver Gasser, John Ronan, Arthur Berger, kc Claffy (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-07506.txt.
Introduction
Hammas Bin Tanveer, Echo Chan, Ricky Mok, and a wider team wrote this research paper, published in 2025. It studies who scans the IPv6 internet, the vast new address space, by setting traps that lure scanners into revealing themselves.
IPv6 is too large to scan blindly, so how scanners pick targets was a mystery. The team built the largest lure network of its kind inside a working internet provider. Over ten months it drew more than six hundred million stray packets from about nineteen hundred networks. The authors sorted the sources, tested which lure features drew the most attention, and pieced together where scanners get their target lists and how they work through them. The findings help defenders anticipate probes in an address space long assumed too big to map.
Catalog record
| CREATOR | Hammas Bin Tanveer, Echo Chan, Ricky K. P. Mok, Sebastian Kappes, Philipp Richter, Oliver Gasser, John Ronan, Arthur Berger, kc Claffy |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 64 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-07506
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Unveiling IPv6 Scanning Dynamics: A Longitudinal Study Using Large Scale Proactive and Passive IPv6 Telescopes Hammas Bin Tanveer, Echo Chan, Ricky K. P. Mok, Sebastian Kappes, Philipp Richter, Oliver Gasser, John Ronan, Arthur Berger, kc Claffy Abstract . We introduce new tools and vantage points to develop and integrate proactive techniques to attract IPv6 scan traffic, thus enabling its analysis. By deploying the largest-ever IPv6 proactive telescope in a production ISP network, we collected over 600M packets of unsolicited traffic from 1.9k Autonomous Systems in 10 months. We characterized the sources of unsolicited traffic, evaluated the effectiveness of five major features across the network stack, and inferred scanners’ sources of target addresses and their strategies. Preprint for a paper accepted at ACM CoNEXT 2025 for publication in the Proceedings of the ACM on Networking. 1. Introduction Internet scanning is a vital tool for researchers and malicious actors alike. Researchers use scanning to characterize network dynamics ( durumeric2013zmap ; bano2018scanning ; markowsky2015scanning ) while malicious actors scan to understand a network’s attack surface. Capturing and analyzing scanning traffic allows network operators and researchers to study scanner behavior and intent, e.g., exploitation of specific vulnerabilities, and consequently build effective defenses against potential malicious traffic. The advent of IPv6 has increased the complexity of the Internet scanning ecosystem. Brute-force scanning of the entire IPv6 address space is practically impossible due to the vastness of address space. Thus, IPv6 scanners must intelligently select targets to increase the likelihood of discovering active addresses. This more sophisticated approach, unlike the brute force approach to IPv4 scanning, renders it harder to detect IPv6 scanning activity. In IPv4, scanning detection leverages darknets : regions of address space that are inactive, i.e. , they neither generate traffic nor host services. Although such dark regions of IPv6 address space are plentiful, they are not as effective for detecting scans, as scanners cannot afford to exhaustively probe such vast inactive address spaces. This reality leaves us with limited visibility into potentially malicious IPv6 scanning activities, hindering our ability to develop methods to secure and protect IPv6 networks, during a time of continuous growth in both IPv6 usage ( google_v6 ) and attempts to exploit it ( CrowdSec32:online ; IPv6thre25:online ) . The potential threat posed by IPv6 scanning has also been recognized within the IETF, which published Internet drafts on operational considerations for IPv6 blocklisting ( draft-ietf-opsec-ipv6-addressing ) and a method for operators to specify end-site prefix lengths of their networks to facilitate sensible IPv6 blocklist entries ( draft-ietf-opsawg-prefix-lengths ) . Figure 1 . Steady increase in IPv6 scan sources per week hitting the CDN. Number of IPv6 addresses (/128s) more than doubled over the two-year window. When aggregated into /64, and /48 subnets, the weekly rate tripled, from \approx 20 to \approx 50-70. Figure 2 . Weekly scan packets (/64 aggregation), grew 100X, from 10-60M to 1B. In early 2022, scan traffic was often dominated by the most active source(s) (dashed lines), by late 2023 scanning traffic comes from a broad range of sources. To validate our assumption that IPv6 scanning traffic is a growing threat, we collaborated with a major Content Delivery Network (CDN) to capture unsolicited IPv6 packets reaching a subset of the CDN’s servers (230,000 machines in over 700 Autonomous Systems (ASes)) from January 1, 2022 to January 1, 2024. Figures 1 and 2 show the steady growth in weekly counts of scanning source IPs, scanning subnets, and scanning packet volume reaching the CDN over this two-year period. 1 1 1 We define a scan as a source hitting at least 100 IPv6 addresses of the CDN, with a maximum packet inter-arrival time of 3,600 seconds, which allows comparability with earlier work that used this timeout for IPv6 scan detection ( richter2022illuminating ) . This earlier work analyzed the sensitivity of the timeout (3600s, 1800s, 900s) and found only marginal differences, with scan detection rates declining by single-digit percentages under shorter thresholds. Given this trade-off, and for a lower bound of 100 probed targets to declare a scan, a timeout of 1 hour was a reasonable choice for our dataset. The main point of the figures is not the number of scans, but rather its significant increase over time. The prefix aggregation accounts for scanners that choose random source addresses within larger prefixes to evade scan detection ( richter2022illuminating ) . (Figure 13 in Appendix C shows a similar growth in number of ASes sourcing scanning traffic). Not only has scanning traffic volume increased two orders of magnitude but it is also more broadly distributed, no longer dominated by one or two scanning sources as it was in early 2022 (see Appendix C for more details). This growing threat motivates our goal in this study: to develop novel and reproducible techniques to attract, capture, and analyze IPv6 scan traffic. This paper describes our three major contributions: 1. New methods and tools to capture IPv6 scan traffic: We developed and implemented novel methods to build proactive telescopes to attract unsolicited IPv6 scanning traffic (§ 3 ). Our approach not only reacts to incoming traffic, but also emits network signals to attract IPv6 scanners. We designed four proactive attraction features: announcing Border Gateway Protocol (BGP) prefixes, registering domain names, issuing Transport Layer Security (TLS) certificates, inserting addresses in IPv6 hitlists. Also, we deployed reactive features–honeypots of different interaction levels to attract IPv6 scanners. To the best of our knowledge, this work integrates the most comprehensive set of features among the prior work we surveyed (§ 2.3 ). In collaboration with a regional Internet Service Provider (ISP), we deployed different combinations of features within subnets of their /32 IPv6 address space, namely honeyprefixes (§ 4.2 ), and compared the results to two passive telescope networks. Our experimental design and deployment strategy allowed us to systematically examine the effect size of individual components in attracting scan activities. We will release the source code of our tools to support reproducibility. 2. Characterization of modern IPv6 scanning: Our proactive techniques induced an increase of three orders of magnitude in unsolicited IPv6 traffic to our previously un-probed honeyprefixes. We stratified deployment of techniques across different subnets to reveal the magnitude and type of IPv6 scanning each technique attracts. Over 90% of observed scanning traffic used ICMP ping, even though the telescopes were responsive to TCP/UDP traffic (§ 5 ). We also found that scanning traffic/strategies differed vastly by source ASN; e.g., Internet Scanner ASNs mostly sent TCP/UDP packets whereas most scanning traffic from Hosting/Cloud providers was ICMP. 3. Implications to IPv6 security: Our findings can help to evaluate and improve the efficacy of network security tools to avoid collateral damage. For example, we found scanners used an entire /30 to send scanning traffic, compared to a /96 for some cloud providers. Awareness of address allocation block size is more critical to safely deploying IPv6 blocklists than in IPv4. 2. Background We provide an overview of the use of network telescopes in capturing unsolicited network packets from darknets, IPv6 Internet scanning techniques, and complexities introduced by IPv6 address allocation and assignment practices to network security tools/techniques. 2.1. Internet Darknets Darknets are regions of unused IP address space that do not emit any network traffic; thus,