Up until recently, Docker servers misconfigured and left exposed online have been historically targeted with cryptocurrency-mining malware, which has helped criminal groups generate huge profits by hijacking someone else’s cloud resources.
However, in a report published this week, security researchers from Trend Micro have discovered what appears to be the first organized and persistent series of attacks against Docker servers that infect misconfigured clusters with DDoS malware.
According to Trend Micro, the two botnets are running versions of the XORDDoS and the Kaiji malware strains. Both malware operations have a long and well-documented history, especially XORDDoS, which has been spotted used in the wild for many years.
However, the two DDoS botnets had usually targeted routers and smart devices, and never complex cloud setups, such as Docker clusters.
“XORDDoS and Kaiji have been known to leverage telnet and SSH for spreading before, so I see Docker as a new vector which increases the potential of the botnet, a green field full of fresh fruit to pick with no immediate competitors,” Pascal Geenens, cybersecurity evangelist at Radware, told ZDNet via email earlier this week.
“Docker containers will typically provide more resources compared to IoT devices, but they typically run in a more secured environment, and it might be hard to impossible for the container to perform DDoS attacks,” Geenens added.
“The unique perspective of IoT devices such as routers and IP cameras is that they have unrestricted access to the internet, but typically with less bandwidth and less horsepower compared to containers in a compromised environment,” the Radware researcher told ZDNet.
“Containers, on the other hand, typically have access to way more resources in terms of memory, CPU, and network, but the network resources might be limited to only one or a few protocols, resulting in a smaller arsenal of DDoS attack vectors supported by those ‘super’ bots.”
However, these limitations don’t usually impact crypto-mining botnets, which only need an open HTTPS channel to the outside world, Geenens said.
But despite the limitations in how a DDoS gang could abuse hacked Docker clusters, Geenens says this won’t stop hackers from attacking this “green field full of fresh fruit to pick” as there are very few vulnerable IoT devices that haven’t been infected already, which has forced hackers to target Docker servers to begin with.
And on a side note, Geenens also told ZDNet that he suspects that DDoS operators are already quite familiar with Docker systems already.
While this is the first time they’re hacking Docker clusters, Geenens believes hackers often use Docker to manage their own attack infrastructure.
“I have no immediate proof, but I’m pretty sure that in the same way as legitimate applications benefit from [Docker’s] automation and agility (DevOps), so will illegal applications.”
The most common source of Docker hacks is the management interface (API) being left exposed online without authentication or being protected by a firewall. For readers looking to secure their servers, that would be a good first thing to check.
In its report, Trend Micro also recommends that server administrators secure their Docker deployments by following a series of basic steps, detailed here.