Distributed denial-of-service attacks targeting large Internet service providers surged in the third quarter of last year as attackers began using a stealthy new tactic to sneak DDoS traffic past detection mechanisms.
An analysis of DDoS data during Q3 2018 by Nexusguard showed attackers trying to overwhelm targeted sites, and even entire ISP -- aka communications provider (CSP) -- networks, by spreading attack traffic across a large number of IP prefixes. Unlike a typical volumetric attack on a single IP address, many of the DDoS campaigns that Nexusguard analyzed involved attackers contaminating legitimate traffic across hundreds of IP addresses with small bits of junk.
The attack traffic within each IP address was small enough to avoid detection by DDoS mitigation tools but big enough to take down a targeted site once converged, Nexusguard said in a report published this week. For example, the average attacks involved just 33.2Mbps of traffic per targeted IP making it hard for service providers to detect and mitigate the traffic.
In total, about 159 autonomous systems - most belonging to service providers - were targeted in "bit-and-piece" attacks in Q3 of 2018. In many of the attacks, hackers appear to have conducted reconnaissance to map out the CSP network and identify critical IP address ranges, Nexusguard said.
The largest attacks involved small bits of DDoS attacks traffic being dispersed to IP addresses across 38 IP prefixes. The maximum number of targeted IP addresses per prefix in these attacks was 252, and the average was 141.
Attacks sizes per IP address ranged from over 300Mbps at the high-end, to just 2.5Mbps at the low-end, with the average at 33.2Mbps. The average attack size per IP prefix was 2.48Gbps. In a worst-case scenario, an attack of this size spread across 38 IP prefixes is potent enough to overwhelm a 10Gbps ISP line, Nexusguard's report noted.
Donny Chong, product director at Nexusguard, says such attacks highlight the need for enterprises to pay closer attention to the DDoS mitigation abilities of their communications service providers.
DDoS detection and mitigation mechanisms that are based on traffic volume thresholds alone are not sufficient in attacks involving very small volumes of attack traffic, he says. Typical anti-DDoS measures like blackholing traffic is not going to work well either, because of the large number IP addresses that are being used to distribute the attack traffic, Chong notes.
CSPs must step up mitigation measures and go beyond the usual threshold and traffic anomaly detection controls, he says.
Of the thousands of DDoS attacks around the world that Nexusguard measured for its study, nearly 66% or two-thirds, were targeted at CSPs. Simple Service Discovery Protocol (SSDP) amplification attacks—a type of DNS amplification attack—increased more than seven-fold in Q3 compared to Q2 2018 because of the increased focus on CSPs.
Because so many DDoS attacks in last year's third-quarter involved small traffic volumes, average attack size decreased 82% year-over-year to 0.97Gbps and by more than 96% compared to Q2 2018.
DDoS in Stealth
The new attacks are a continued trend of threat actors evolving DDoS attacks to sneak them past enterprise security controls. Despite the smaller average sizes that Nexusguard reported, DDoS attacks have generally gotten bigger, more sophisticated and multi-faceted in recent years. Many involve multiple attack vectors and, increasingly, services are becoming available that carry out attacks on request or make available tools that allow almost anyone to launch a DDoS attack.
"In 2019, we anticipate more attackers crawling out of the woodwork to offer their services to the highest bidder," NETSCOUT said in a blog earlier this month. Many of these services will provide DIY DDoS tools that lower the bar for attackers, the blog noted.
"While these tools are not necessarily new to the scene, the ease of access, quick iteration at including new attack types, and a broader range of international customers will result in lots of amateur cybercriminals getting hold of destructive malware," the company said.
Jai Vijayan is a seasoned technology reporter with over 20 years of experience in IT trade journalism. He was most recently a Senior Editor at Computerworld, where he covered information security and data privacy issues for the publication. Over the course of his 20-year ... View Full Bio