DNS Internet Papers Publications Uncategorized

new paper “Defending Root DNS Servers Against DDoS Using Layered Defenses” at COMSNETS 2023

Our paper titled “Defending Root DNS Servers Against DDoS Using Layered Defenses” will appear at COMSNETS 2023 in January 2023. In this work, by ASM Rizvi, Jelena Mirkovic, John Heidemann, Wes Hardaker, and Robert Story, we design an automated system named DDIDD with multiple filters to handle an ongoing DDoS attack on a DNS root server. We evaluated ten real-world attack events on B-root and showed DDIDD could successfully mitigate these attack events. We released the datasets for these attack events on our dataset webpage (dataset names starting with B_Root_Anomaly).

From the abstract:

Distributed Denial-of-Service (DDoS) attacks exhaust resources, leaving a server unavailable to legitimate clients. The Domain Name System (DNS) is a frequent target of DDoS attacks. Since DNS is a critical infrastructure service, protecting it from DoS is imperative. Many prior approaches have focused on specific filters or anti-spoofing techniques to protect generic services. DNS root nameservers are more challenging to protect, since they use fixed IP addresses, serve very diverse clients and requests, receive predominantly UDP traffic that can be spoofed, and must guarantee high quality of service. In this paper we propose a layered DDoS defense for DNS root nameservers. Our defense uses a library of defensive filters, which can be optimized for different attack types, with different levels of selectivity. We further propose a method that automatically and continuously evaluates and selects the best combination of filters throughout the attack. We show that this layered defense approach provides exceptional protection against all attack types using traces of real attacks from a DNS root nameserver. Our automated system can select the best defense within seconds and quickly reduce the traffic to the server within a manageable range while keeping collateral damage lower than 2%. We can handle millions of filtering rules without noticeable operational overhead.

This work is partially supported by the National Science
Foundation (grant NSF OAC-1739034) and DHS HSARPA
Cyber Security Division (grant SHQDC-17-R-B0004-TTA.02-
0006-I), in collaboration with NWO.


USC/Viterbi and ISI news about “Anycast Agility” paper

USC Viterbi and ISI both posted a news article about our paper “Anycast Agility: Network Playbooks to Fight DDoS”.

Please see our blog entry for the abstract and the full technical paper for the real details, but their posts are very accessible. And with the hacker in the hoodie, you know it’s serious :-)


new paper “Differences in Monitoring the DNS Root Over IPv4 and IPv6” to appear at the IEEE National Symposium for NSF REU Research in Data Science, Systems, and Security

On December 15, 2022, Tarang Saluja will present the paper “Differences in Monitoring the DNS Root Over IPv4 and IPv6” (by Tarang Saluja, John Heidemann, and Yuri Pradkin) at the IEEE National Symposium for NSF REU Research in Data Science, Systems, and Security.

From the abstract:

Figure 9 from [Saluja22a], showing fraction of query failures in RIPE Atlas after we remove observers that are islands (unable to reach any of the 13 DNS root identifiers). Blue is IPv4, red is IPv6, with data for each of the 13 DNS root identifiers. We believe this data is a better representation of what people expect to see than Atlas results that include these “broken” observers.

The Domain Name System (DNS) is an essential service for the Internet which maps host names to IP addresses. The DNS Root Sever System operates the top of this namespace. RIPE Atlas observes DNS from more than 11k vantage points (VPs) around the world, reporting the reliability of the DNS Root Server System in DNSmon. DNSmon shows that loss rates for queries to the DNS Root are nearly 10% for IPv6, much higher than the approximately 2% loss seen for IPv4. Although IPv6 is “new,” as an operational protocol available to a third of Internet users, it ought to be just as reliable as IPv4. We examine this difference at a finer granularity by investigating loss at individual VPs. We confirm that specific VPs are the source of this difference and identify two root causes: VP islands with routing problems at the edge which leave them unable to access IPv6 outside their LAN, and VP peninsulas which indicate routing problems in the core of the network. These problems account for most of the loss and nearly all of the difference between IPv4 and IPv6 query loss rates. Islands account for most of the loss (half of IPv4 failures and 5/6ths of IPv6 failures), and we suggest these measurement devices should be filtered out to get a more accurate picture of loss rates. Peninsulas account for the main differences between root identifiers, suggesting routing disagreements root operators need to address. We believe that filtering out both of these known problems provides a better measure of underlying network anomalies and loss and will result in more actionable alerts.

Original data from this paper is available from RIPE Atlas (measurement ids are in the paper). We are publishing new results daily on our website (from the RIPE data).

This work was done while Tarang was on his Summer 2022 undergraduate research internship at USC/ISI, with support from NSF grant 2051101 (PI: Jelena Mirkovich). John Heidemann and Yuri Pradkin’s work is supported by NSF through the EIEIO project (CNS-2007106). We thank Guillermo Baltra for his work on islands and peninsulas, as seen in his arXiv report.

Outages Presentations Publications Uncategorized

new poster “Internet Outage Detection Using Passive Analysis” at ACM IMC 2022

Asma Enayet will present her poster “Internet Outage Detection Using Passive Analysis” by Asma Enayet and John Heidemann at ACM Internet Measurement Conference, Nice, France from October 25-27th, 2022.

We expect the ACM poster abstract (without the poster) to appear at in October 2022.

We are making a report available now with the poster abstract and poster at as a pre-print.

From the abstract:

Outages from natural disasters, political events, software or hardware issues, and human error place a huge cost on e-commerce ($66k per minute at Amazon). While several existing systems detect Internet outages, these systems are often too inflexible, with fixed parameters across the whole internet with CUSUM-like change detection. We instead propose a system using passive data, to cover both IPv4 and IPv6, customizing parameters for each block to optimize the performance of our Bayesian inference model. Our poster describes our three contributions: First, we show how customizing parameters allows us often to detect outages that are at both fine timescales (5 minutes) and fine spatial resolutions (/24 IPv4 and /48 IPv6 blocks). Our second contribution is to show that, by tuning parameters differently for different blocks, we can scale back temporal precision to cover more challenging blocks. Finally, we show our approach extends to IPv6 and provides the first reports of IPv6 outages.

IPv6 Coverage: our source of passive data (B-Root) is incomplete, but it provides similar coverage in both IPv4 and IPv6.
IPv6 Outages: Outage rate for IPv6 (12%) is greater than for IPv4 (5.5%) —IPv6 reliability can improve.

This work was supported by NSF grant CNS-2007106 (EIEIO).


Hurricane Ian Seen through Internet Outages in Florida

We’ve been watching Hurricane Ian move across the Gulf of Mexico as it approached Florida. It looks like it made landfall about 3pm EDT (19:00Z).

Our Trinocular Internet outage detection showed the first residential Internet outages starting around 2:20pm EDT (18:22Z) and they go up from there. Our most recent data (as of this post at 7:45pm EDT, or 23:45Z) shows outages from about 75 minutes ago at 6:15pm EDT (22:13Z) with network outages over 50% in most of the western peninsula.

Trinocular outages in Florida at 6:13pm EDT (22:13Z). Circle area is proportional to the number of networks that are out in each 0.5×0.5 degree geographic grid cell, the color is the percentage of networks that are out.

We hope folks stay inside and safe!


congratulations to Tarang Saluja for his summer undergraduate research internship

Tarang Saluja completed his summer undergraduate research internship at ISI this summer, working with John Heidemann and Yuri Pradkin on his project “Differences in Monitoring the DNS Root Over IPv4 and IPv6″.

In his project, Tarang examined RIPE Atlas’s DNSmon, a measurement system that monitors the Root Server System. DNSmon examines both IPv4 and IPv6, and its IPv6 reports show query loss rates that are consistently higher than IPv4, often 4-6% IPv6 loss vs. no or 2% IPv4 loss. Prior results by researchers at RIPE suggested these differences were due to problems at specific Atlas Vantage Points (VPs, also called Atlas Probes).

Tarang Saluja describing his research to an ISI researcher, at the ISI REU Poster Session on 2022-08-01.

Building on the Guillero Baltra’s studies of partial connectivity in the Internet, Tarang classified Atlas VPs with problems as islands and peninsulas. Islands think they are on IPv6, but cannot reach any of the 13 Root DNS “letters” over IPv6, indicating that the VP has a local network configuration problem. Peninsulas can reach some letters, but not others, indicating a routing problem somewhere in the core of the Internet.

Tarang’s work is important because these observations allow lead to potential solutions. Islands suggest VPs that do not support IPv6 and so should not be used for monitoring. Peninsulas point to IPv6 routing problems that need to be addressed by ISPs. Setting VPs with these problems aside provides a more accurate view of what IPv6 should be, and allows us to use DNSmon to detect more subtle problems. Together, his work points the way to improving IPv6 for everyone and improving Root DNS access over IPv6.

Tarang’s work was part of the ISI Research Experiences for Undergraduates program at USC/ISI. We thank Jelena Mirkovic (PI) for coordinating another year of this great program, and NSF for support through award #2051101.

Technical Report

new technical report: Having your Privacy Cake and Eating it Too: Platform-supported Auditing of Social Media Algorithms for Public Interest

We have released a new technical report: “Having your Privacy Cake and Eating it Too: Platform-supported Auditing of Social Media Algorithms for Public Interest”, available at

From the abstract:

Legislations have been proposed in both the U.S. and the E.U. that mandate auditing of social media algorithms by external researchers. But auditing at scale risks disclosure of users’ private data and platforms’ proprietary algorithms, and thus far there has been no concrete technical proposal that can provide such auditing. Our goal is to propose a new method for platform-supported auditing that can meet the goals of the proposed legislations. The first contribution of our work is to enumerate these challenges and the limitations of existing auditing methods to implement these policies at scale. Second, we suggest that limited, privileged access to relevance estimators is the key to enabling generalizable platform-supported auditing of social media platforms by external researchers. Third, we show platform-supported auditing need not risk user privacy nor disclosure of platforms’ business interests by proposing an auditing framework that protects against these risks. For a particular fairness metric, we show that ensuring privacy imposes only a small constant factor increase (6.34× as an upper bound, and 4× for typical parameters) in the number of samples required for accurate auditing. Our technical contributions, combined with ongoing legal and policy efforts, can enable public oversight into how social media platforms affect individuals and society by moving past the privacy-vs-transparency hurdle.

High-level overview of our proposed platform-supported framework for auditing relevance estimators while protecting the privacy of audit participants and the business interests of platforms.

This technical report is a joint work of Basileal Imana from USC, Aleksandra Korolova from Princeton University, and John Heidemann from USC/ISI.


Large Canadian Internet Outage

News reports (for example, at the Verge and Slashdot) mention a large outage in Rogers, a major Canadian telecommunications provider.

We see lots of evidence for this in our Internet outage detection system.

It’s big! Maybe 30% of Toronto and southern Ontario networks, plus a lot of outages in New Brunswick.


Internet outages in Ontario, Canada. The largest circle represents about 6500 /24 network blocks down near Toronto, about 30% of the /24 blocks in that area. See details on our outage website.

New Brunswick:

Internet outages in New Brunswick, Canada. The largest circle here represents 196 /24 network blocks down near Moncton, more than 45% of the /24 blocks there. The red circles are areas where most or all network blocks are currently out. See details on our outage website.

An update: Newfoundland also sees a lot of outages. Quebec looks in pretty good shape, though.

And it’s lasting a long time. It looks like it started at 5am Eastern time (2022-07-08t09:00Z), it it has lasted 9.5 hours so far!

We wish Rogers personnel and our Canadian neighbors the best.

Update at 2022-07-09t06:15Z (2:15am Eastern time): Toronto is doing much better, with “only” 10% of blocks unreachable (22808 of 21.5k in the 43.8N,79.3W 0.5 grid cell). New Brunswick and Newfoundland still look the same, with outages in about 50% of blocks.

Update at 2022-07-09t21:10Z (5:10pm Eastern time): It looks like many Rogers networks recovered at 2022-07-09t05:15Z (1:15am Eastern time). This includes all of New Brunswick and Newfoundland and most of Ontario. Trinocular has about a one-hour delay while it computes results, so I did not see this result when I checked in the prior update–I needed to wait 15 minutes more.


Internet Outages Timelines and Events in 2022

We recently added timeline support to our Outage World map–clicking on an outage bubble pops up a window with a sparkline (a small graph) showing maximum outages on each data for the current quarter, and clicking on the “daily timeline” tab shows outages for the current 24 hours. These graphs help provide context for how long an outage lasts, and if there were other outages the same quarter.

As an example, here is a major outage effecting most of central and southern Mexico on 2022-01-05. The timeline of Mexico City shows how unusual this outage was:

Some other big outages in 2022 include this big outage in Italy on April 27 from 18:00 to 23:59:

and in southwest Florida on April 24 at 3:15pm Eastern Time (that’s 2022-04-24t19:15Z) that was confirmed as a fiber cut:

Thanks to Erica Stutz for adding timelines to the outage code (as a follow on to her work on Covid-19 Work-from-Home visualization) and to Yuri Pradkin for spotting these events.


new paper “Old but Gold: Prospecting TCP to Engineer and Live Monitor DNS Anycast” Awarded Best Paper at the Passive and Active Measurement Conference

On March 29, 2022 the paper “Old but Gold: Prospecting TCP to Engineer and Live Monitor DNS Anycast” by Giovane C. M. Moura, John Heidemann, Wes Hardaker, Pithayuth Charnsethikul, Jeroen Bulten, João M. Ceron, and Cristian Hesselman appeared that the 2022 Passive and Active Measurement Conference. We’re happy that it was awarded Best Paper for this year’s conference!

From the abstract:

Google latency for .nl before (left red area) and after (middle green area) DNS polarization was corrected. Polarization was detected with ENTRADA using the work from this paper.

DNS latency is a concern for many service operators: CDNs exist to reduce service latency to end-users but must rely on global DNS for reachability and load-balancing. Today, DNS latency is monitored by active probing from distributed platforms like RIPE Atlas, with Verfploeter, or with commercial services. While Atlas coverage is wide, its 10k sites see only a fraction of the Internet. In this paper we show that passive observation of TCP handshakes can measure live DNS latency, continuously, providing good coverage of current clients of the service. Estimating RTT from TCP is an old idea, but its application to DNS has not previously been studied carefully. We show that there is sufficient TCP DNS traffic today to provide good operational coverage (particularly of IPv6), and very good temporal coverage (better than existing approaches), enabling near-real time evaluation of DNS latency from real clients. We also show that DNS servers can optionally solicit TCP to broaden coverage. We quantify coverage and show that estimates of DNS latency from TCP is consistent with UDP latency. Our approach finds previously unknown, real problems: DNS polarization is a new problem where a hypergiant sends global traffic to one anycast site rather than taking advantage of the global anycast deployment. Correcting polarization in Google DNS cut its latency from 100ms to 10ms; and from Microsoft Azure cut latency from 90ms to 20ms. We also show other instances of routing problems that add 100-200ms latency. Finally, real-time use of our approach for a European country-level domain has helped detect and correct a BGP routing misconfiguration that detoured European traffic to Australia. We have integrated our approach into several open source tools: Entrada, our open source data warehouse for DNS, a monitoring tool (ANTS), which has been operational for the last 2 years on a country-level top-level domain, and a DNS anonymization tool in use at a root server since March 2021.

The tools we developed in this paper are freely available, including patches to Knot, improvements to dnsanon, improvements to ENTRADA, and the new tool Anteater. Unfortunately data from the paper was from operational DNS systems and so cannot be shared due to privacy concerns.

This paper was made in part through DHS HSARPA Cyber Security Division via contract number HSHQDC-17-R-B0004-TTA.02-0006-I (PAADDOS) and by NWO, NSF CNS-1925737 (DIINER), and the Conconrdia Project, an European Union’s Horizon 2020 Research and Innovation program under Grant Agreement No 830927.