Attackers abuse legacy routing protocol to amplify DDoS attacks
DDoS attacks can be amplified through SOHO routers that still use the Routing Information Protocol version 1, researchers warn.
By Lucian Constantin
Servers could be haunted by a ghost from the 1980s, as hackers have started abusing an obsolete routing protocol to launch distributed denial-of-service attacks.
DDoS attacks observed in May by the research team at Akamai abused home and small business (SOHO) routers that still support Routing Information Protocol version 1 (RIPv1). This protocol is designed to allow routers on small networks to exchange information about routes.
RIPv1 was first introduced in 1988 and was retired as an Internet standard in 1996 due to multiple deficiencies, including lack of authentication. These were addressed in RIP version 2, which is still in use today.
In the DDoS attacks seen by Akamai, which peaked at 12.8 gigabits per second, the attackers used about 500 SOHO routers that are still configured for RIPv1 in order to reflect and amplify their malicious traffic.
DDoS reflection is a technique that can be used to hide the real source of the attack, while amplification allows the attackers to increase the amount of traffic they can generate.
RIP allows a router to ask other routers for information stored in their routing tables. The problem is that the source IP (Internet Protocol) address of such a request can be spoofed, so the responding routers can be tricked to send their information to an IP address chosen by attackers—like the IP address of an intended victim.
This is a reflection attack because the victim will receive unsolicited traffic from abused routers, not directly from systems controlled by the attackers.
But there’s another important aspect to this technique: A typical RIPv1 request is 24-byte in size, but if the responses generated by abused routers are larger than that, attackers can generate more traffic they could otherwise do with the bandwidth at their disposal.
In the attacks observed by Akamai, the abused routers responded with multiple 504-byte payloads—in some cases 10—for every 24-byte query, achieving a 13,000 percent amplification.
Other protocols can also be exploited for DDoS reflection and amplification if servers are not configured correctly, including DNS (Domain Name System), mDNS (multicast DNS), NTP (Network Time Protocol) and SNMP (Simple Network Management Protocol).
The Akamai team scanned the Internet and found 53,693 devices that could be used for DDoS reflection using the RIPv1 protocol. Most of them were home and small business routers.
The researchers were able to determine the device make and model for more than 20,000 of them, because they also had their Web-based management interfaces exposed to the Internet.
Around 19,000 were Netopia 3000 and 2000 series DSL routers distributed by ISPs, primarily from the U.S., to their customers. AT&T had the largest concentration of these devices on its network—around 10,000—followed by BellSouth and MegaPath, each with 4,000.
More than 4,000 of the RIPv1 devices found by Akamai were ZTE ZXV10 ADSL modems and a few hundred were TP-Link TD-8xxx series routers.
While all of these devices can be used for DDoS reflection, not all of them are suitable for amplification. Many respond to RIPv1 queries with a single route, but the researchers identified 24,212 devices that offered at least an 83 percent amplification rate.
To avoid falling victim to RIPv1-based attacks, server owners should use access control lists to restrict Internet traffic on UDP source port 520, the Akamai researchers said in their report. Meanwhile, the owners of RIPv1-enabled devices should switch to RIPv2, restrict the protocol’s use to the internal network only or, if neither of those options is viable, use access control lists to restrict RIPv1 traffic only to neighboring routers.
Note: When you purchase something after clicking links in our articles, we may earn a small commission. Read ouraffiliate link policyfor more details.