It's difficult for security professionals to communicate just how relentless their work is, especially when resurgent threats emerge. If you arent dealing with the "whack-a-mole" nature of the security landscape which has threats emerging in new and different forms, it may be easy to think that once a threat is dealt it that it will stay dealt with.
CenturyLink belies that view with a new report that it issued last week. CenturyLink Threat Research Labs has been tracking an IoT botnet called "TheMoon," which was first observed in 2014.
This botnet exploits target broadband modems or routers that were developed by companies like Linksys, ASUS, MikroTik and D-Link. The most recent exploit (which was added in May 2018) targets GPON routers. Most of the exploits will target vulnerabilities in IoT web applications which are usually running on port 8080.
Initially, the botnet was a DDoS implementer, but has not shown up in this role for a while.
What CenturyLink has found is that TheMoon has pivoted itself to be a proxy botnet that has a completely different threat model associated with itself than it previously did.
It enabled this new behavior with specific architecture changes. CenturyLink found that the main binary uses three different ports for command and control communication. The first is for initial registration when the binary starts executing, one is for command and control communication and the last will download additional payloads to run.
These ports vary between binaries and architecture types. For example, MIPS binaries use port 5784 for registration, 5184 for command and control and 4584 for downloading payloads while ARM binaries use ports 5732, 5132 and 4532, respectively.
The separation of function implied by the differing ports is what allows TheMoon to be rented out by others. CenturyLink found that a new, previously undisclosed module running on TheMoon would turn the device into a SOCKS5 proxy for others to use.
The proxy port was found by CenturyLink to be a randomly chosen port above 10,000 and was observed by them to change multiple times per day.
Observing the botnet traffic over time let CenturyLink determine that the botnet was involved in serving out embedded YouTube videos for fraud purposes. This was possible because the operators that were renting the bot left an exposed port which gave CenturyLink the log data from the servers.
They did notice that the geolocation field of the profile for a device making requests did not correspond with the geolocation of the proxy IP that was supposedly making the request.
CenturyLink thinks that this won't be the last incarnation of this kind of threat. In the report they say that, "There is also a substantial market for proxy botnets targeting broadband networks to route traffic for attacks like credential brute forcing and ad fraud. The always-on nature of IoT devices and the ability to masquerade as normal home users make broadband networks prime targets for these types of attacks."
An old threat mutates into a new threat, driven by the economics of the situation. Such is the security process. It's never finished, and must always be aware of change.
— Larry Loeb has written for many of the last century's major "dead tree" computer magazines, having been, among other things, a consulting editor for BYTE magazine and senior editor for the launch of WebWeek.