5 Items to Monitor to Detect DDoS Attacks
DKosig/istockphoto.com
Government organizations have been a steady target of distributed denial-of-service attacks.
Distributed denial-of-service attacks have steadily increased over the last 18 months, despite being overshadowed by malware attacks and ransomware.
Kaspersky Labs estimated 24% more DDoS attacks in 2020 than in 2019 and F5 Networks found that DDoS attacks increased by 55% between January 2020 and March 2021. Notable DDoS attacks targeted Amazon Web Services in February 2020, GitHub in February 2018 and Google in 2017.
Government organizations have been a steady target of DDoS attacks; F5 Networks and IBM X-Force both found that government agencies were the sixth-most targeted vertical in 2020. Since these attacks involve hijacking or abusing network protocols, one way for government IT departments to detect them is to monitor certain types of network traffic.
Here are five network packet types and protocols commonly abused in DDoS attacks and guidance on how best to monitor them.
- Transmission Control Protocol Synchronize (TCP-SYN): This type of attack uses a flood of TCP SYN (the initial packets from client to server to establish a session) packets to consume enough server resources to render the system unresponsive to all legitimate traffic. Monitoring for significant increases in TCP SYN packets will show when a flood is incoming.
- Domain Name System (DNS): Monitoring DNS activity is essential to identifying early signs of a DNS flood DDoS attack. DNS uses two types of packets: DNS Request and DNS Response. To detect attacks, both types need to be independently monitored; in the case of a DDoS attack, the number of DNS Request packets will be considerably higher than the number of DNS Response packets and SecOps should be alerted when the ratio goes beyond a rational number.
- Application Flooding: Application attacks, such as an HTTP flood, target layer 7 in the OSI model rather than network infrastructure like DNS. They’re effective because they can consume both server and network resources, less traffic is required to cause disruption, and it’s difficult for defenders to tell the difference between attack traffic and legitimate traffic.
- User Datagram Protocol (UDP): UDP fragmentation, amplification and flooding attacks were the most common type of DDoS attack in 2020 according to F5 Networks. In these attacks, an attacker sends a valid UDP request packet listing the target’s IP as the UDP source IP address to a server, which sends back a much larger response to the target’s IP. By using UDP packets over 1,500 bytes in size, the attacker can force the packet to be fragmented (as ethernet MTU is 1,500 bytes). These techniques amplify the volume of traffic directed against the victim (hence the “flooding”). Monitoring for UDP traffic outside of normal levels, especially the protocols listed below that are ripe for abuse, will allow NetOps and IT teams to see when a UDP attack may be occurring.
- Internet Control Message Protocol (ICMP): ICMP is another transmission protocol that is susceptible to storms, so monitoring the overall throughput and count of ICMP packets will provide early warning of internal or external issues. ICMP Address Mask requests and ICMP Type 9 and Type 10 protocols can also be used to launch DoS and MitM attacks. To mitigate these attacks, IT should disallow ICMP route discovery and use digital signatures to block all type 9 and type 10 ICMP packets.
While monitoring these systems and protocols is an important step, network visibility in government agencies does have differences compared to the private sector. For instance, many agencies are required to store packet data for certain intervals (10 days, 30 days, etc.), requiring that the chosen monitoring solution and packet capture (PCAP) storage can scale accordingly. Even a week’s worth of packet data can take up gigabytes of space, so the amount of storage required for this can be considerable—and needs to increase to keep pace with network usage gains. Moreover, agencies looking to upgrade network infrastructure to 40 Gbps or 100 Gbps speeds will need to be even more mindful of scaling issues, as lossless packet capture and analysis (which is essential for maintaining security) at these speeds is technically challenging. As a result, many products don’t yet support lossless collection above 40Gbps.
Appropriate network visibility, security delivery (specifically replication forwarding services) and packet capture storage capability is critical to maintaining performance, security and providing PCAP forensics of government infrastructure. Having the right tools and looking at the right protocols will allow agencies to stay ahead of the latest threats to ongoing operations.
Vince Hill is a technical marketing engineer at cPacket Networks.