Resources on the Internet allow constant communication and data sharing between Internet users. While these resources keep vital information flowing, cybercriminals can easily compromise and abuse them, using them as a platform for fraud and misuse. Every day, we observe millions of internet-connected resources are being abused in criminal activities, ranging from poorly-configured Internet of Things (IoT) devices recruited into flooding legitimate services’ networks with unwanted Internet traffic or compromising legitimate websites to distribute malicious software that is designed to prevent access to victim’s data or device until a ransom has been paid to the attacker. The Internet's decentralized architecture necessitates that defenders must voluntarily collaborate to combat cybercrime. While mandatory efforts may be necessary in some circumstances, the bulk of incident response will remain based on voluntary actions among thousands of Internet intermediaries, researchers and resource owners. These voluntary actions typically take the form of one party sending security notifications to another about potential security issues and asking them to act against it. Security notifications are intended to support and promote a wide range of feasible efforts, which aim to detect and mitigate millions of daily incidents and remediate underlying conditions. Despite its importance, voluntary action remains a poorly understood and significantly less investigated component of the fight against cybercrime. All of this puts a premium on understanding how voluntary cyber-defense efforts prove to be the most effective in remediating security issues.@en

With the rise of IoT botnets, the remediation of infected devices has become a critical task. As over 87% of these devices reside in broadband networks, this task will fall primarily to consumers and the Internet Service Providers. We present the first empirical study of IoT malware cleanup in the wild -- more specifically, of removing Mirai infections in the network of a medium-sized ISP. To measure remediation rates, we combine data from an observational study and a randomized controlled trial involving 220 consumers who suffered a Mirai infection together with data from honeypots and darknets. We find that quarantining and notifying infected customers via a walled garden, a best practice from ISP botnet mitigation for conventional malware, remediates 92% of the infections within 14 days. Email-only notifications have no observable impact compared to a control group where no notifications were sent. We also measure surprisingly high natural remediation rates of 58-74% for this control group and for two reference networks where users were also not notified. Even more surprising, reinfection rates are low. Only 5% of the customers who remediated suffered another infection in the five months after our first study. This stands in contrast to our lab tests, which observed reinfection of real IoT devices within minutes -- a discrepancy for which we explore various different possible explanations, but find no satisfactory answer. We gather data on customer experiences and actions via 76 phone interviews and the communications logs of the ISP. Remediation succeeds even though many users are operating from the wrong mental model -- e.g., they run anti-virus software on their PC to solve the infection of an IoT device. While quarantining infected devices is clearly highly effective, future work will have to resolve several remaining mysteries. Furthermore, it will be hard to scale up the walled garden solution because of the weak incentives of the ISPs.@en

Mechanisms for large-scale vulnerability notifications have been confronted with disappointing remediation rates. It has proven difficult to reach the relevant party and, once reached, to incentivize them to act. We present the first empirical study of a potentially more effective mechanism: Quarantining the vulnerable resource until it is remediated. We have measured the remediation rates achieved by a medium-sized ISP for 1, 688 retail customers running open DNS resolvers or Multicast DNS services. These servers can be abused in UDP-based amplification attacks. We assess the effectiveness of quarantining by comparing remediation with two other groups: One group which was notified but not quarantined and another group where no action was taken. We find very high remediation rates for the quarantined users, 87%, even though they can self-release from the quarantine environment. Of those who received the email-only notification, 76% remediated. Surprisingly, over half of the customers who were not notified at all also remediated, though this is tied to the fact that many observations of vulnerable servers are transient. All in all, quarantining appears more effective than other notification and remediation mechanisms, but it is also clear that it can not be deployed as a general solution for Internet-wide notifications.

@en

As large-scale vulnerability detection becomes more feasible, it also increases the urgency to find effective largescale notification mechanisms to inform the affected parties. Researchers, CERTs, security companies and other organizations with vulnerability data have a variety of options to identify, contact and communicate with the actors responsible for the affected system or service. A lot of things can – and do – go wrong. It might be impossible to identify the appropriate recipient of the notification, the message might not be trusted by the recipient, it might be overlooked or ignored or misunderstood. Such problems multiply as the volume of notifications increases. In this paper, we undertake several large-scale notification campaigns for a vulnerable configuration of authoritative nameservers. We investigate three issues: What is the most effective way to reach the affected parties? What communication path mobilizes the strongest incentive for remediation? And finally, what is the impact of providing recipients a mechanism to actively demonstrate the vulnerability for their own system, rather than sending them the standard static notification message. We find that retrieving contact information at scale is highly problematic, though there are different degrees of failure for different mechanisms. For those parties who are reached, notification significantly increases remediation rates. Reaching out to nameserver operators directly had better results than going via their customers, the domain owners. While the latter, in principle, have a stronger incentive to care and their request for remediation would trigger the commercial incentive of the operator to keep its customers happy, this communication path turned out to have slightly worse remediation rates. Finally, we find no evidence that vulnerability demonstrations did better than static messages. In fact, few recipients engaged with the demonstration website. @en

Cybercriminal activity has exploded in the past decade, with diverse threats ranging from phishing attacks to botnets and drive-by-downloads afflicting millions of computers worldwide. In response, a volunteer defense has emerged, led by security companies, infrastructure operators, and vigilantes. This reactionary force does not concern itself with making proactive upgrades to the cyber infrastructure. Instead, it operates on the front lines by remediating infections as they appear. We construct a model of the abuse reporting infrastructure in order to explain how voluntary action against cybercrime functions today, in hopes of improving our understanding of what works and how to make remediation more effective in the future. We examine the incentives to participate among data contributors, affected resource owners, and intermediaries. Finally, we present a series of key attributes that differ among voluntary actions to investigate further through experimentation, pointing toward a research agenda that could establish causality between interventions and outcomes.@en

Motivation: Participants on the front lines of abuse reporting have a variety of options to notify intermediaries and resource owners about abuse of their systems and services. These can include emails to personal messages to blacklists to machine-generated feeds. Recipients of these reports have to voluntarily act on this information. We know remarkably little about the factors that drive higher response rates to abuse reports. One such factor is the reputation of the sender. In this article, we present the first randomized controlled experiment into sender reputation. We used a private datafeed of Asprox-infected websites to issue notifications from three senders with different reputations: an individual, a university and an established anti-malware organization.

Results: We find that our detailed abuse reports significantly increase cleanup rates. Surprisingly, we find no evidence that sender reputation improves cleanup. We do see that the evasiveness of the attacker in hiding compromise can substantially hamper cleanup efforts. Furthermore, we find that the minority of hosting providers who viewed our cleanup advice webpage were much more likely to remediate infections than those who did not, but that website owners who viewed the advice fared no better.@en