Fbot re-emerged, the backstage

Mirai Fbot is back with the stronger infection speed, “Currently FBOT’s infection speed is about 100 nodes per day.. Don’t trust me, trust the number”, said the researcher who figured the come-back…”

In the previous month we covered the mystery behind the Mirai botnet variant dubbed as “Fbot” after the threat’s latest encryption was disclosed by security researcher unixfreaxjp of the MalwareMustDie team. The article  we published ended with the question “Would it mean that the coder of Fbot is abandoning his botnet after all of this time?”

Fbot is one of Mirai’s variants, and Mirai is the Linux malware that originally was detected in August, 2016 by the same team who wrote the analysis. On the boom of Mirai source code leaks, followed by the sharing of its source code openly in the Github by a security researcher, a lot of young hackers involved in the circle of “DDoS criminal ecosystem”, who had been actively using IoT devices for the DDoS purpose were racing to use Mirai to their better DDoS botnet platforms.

Fbot historically has been a part of Mirai variant for some years by now, leveraging the leak source code of other Mirai variants made by different coders and improving them to be better than other rival IoT botnets, obviously to make more infection hits

In this article we successfully interviewed unixfreaxjp,, the researcher who analyzed this re-emerging threat. When he was asked about how he thinks about the FBot re-emerging, he commented “..although we are all hoping that the bad actors will stop their malicious activities seeing that the new propagation had stopped for a month during our monitoring, to only realizing that the actors were only actually put the botnet into sleep and rebuilt their botnet technology to be better in evasion and propagation..”

About how recent Fbot was started to come into his attention, he explained “..basically we monitored all Linux threats non-stop, not only the ones aiming IoT, but all of them, to seek how the Linux threat landscape situated recently,  what technology is used mostly now, and to understand their  trends. We dealt with a lot of malware and their threat actors for these past seven years, and any threat can come-back again as long as the motivation, bad adversaries  and weak target are still laying around, And Fbot is a good example for the case..”

The speed of new variant of Fbot has been reported as super fast, when we came into this topic he replied “..Fbot owned significant amount of botnet before, like thousands IoT devices, they can go back to previous infection volume easy due to the circumstance for spreading the infection is on adversaries side, I think that is what the malware coder and bot herder is expecting from this wave as his first stage. However he learned from previous cases too, by improving the evasion technique using the hexstring-push method to drop the loader into the IoT”

What is the “hexstring-push” method mentioned in the post? Is it a new trick in IoT infection? Could you explain more of it? “..Well, a “hexstring-push” is just a term I made, since I don’t know how to call this techniques,  it’s when an actor is sending a textual-basis data which is actually a binary that can be saved or injected and to be executed in the targeted machine. This is not a new matter since exploitation development is using this concept for so long, and recent post-exploitation frameworks  aiming multiple operating systems are utilizing the same concept too. The adversary is in an experimental stage by using this concept to pivot against commonly known download commands that can block their infection..”

You wrote that there are embedded binaries inside the Mirai FBot, are they the same as the “hex-pushed” ones? What is their purpose? “..Good questions, those embedded ELF are Mirai Fbot downloaders, they call it as “loader”, and yes, in the ones I analyzed it is coded as same as the “hexstring” one. The purpose of this “loader” is to download Fbot from the C2 servers to the successfully compromised IoT device.”

You have just mentioned about C2 (Command and Control) servers, how many are they? And who owns them? “..Yes,since you asked about this, I just updated our report with the C2 servers used, it is explained as per below table. The alleged owners are the threat actor related individuals, that are possible a hands out environment through many hands.”

The published analysis post contains a lot of information of the infected IoT devices and the record of the infection is clearly shown to show all of us the speed of infection. The infection that has been started to be detected on February 22, 2020 as 582 nodes has become 1,086 nodes within 3days, and it was reportedly Mirai Fbot to have the infection rate of 100 nodes per day.

At this point we asked about the handling of fast Fbot infection: “How do you propose to stop Fbot botnet from its fast spreading, knowing that the infection speed is high?”, and the reply is  “..Very good question. As usual in IoT threat, the actors want their botnet to spread wide under the radar and owning as much IoT as they can get, and stay maintaining the ownership of those IoT until they are ready to do some action as a service for several bad activities like DDoS, traffic cushion for further hacks, malware proxy , and so on. Fbot is not different. Lucky that we disclosed the come-back wave early, so hopefully we had the awareness in place, that’s the first important step. The next step is, no one can fight this type of threat alone, we are all have different limitation, but now security communities are all in a form of coordination to work together to mitigate this type of the threat, by spotting new infection techniques, recording new C2 for blocking purpose, patching the vulnerability the botnet aims to infect, informing CSIRT for services to be more aware of the threat, and applying better IoT policies. Hopefully these steps will suppress the propagation speed of the infection.”

In their post it is clearly written there are countries that are having faster infection than others. When asked about this matter, this is what unixfreaxjp replied: “IoT devices are varied in many countries, not only about variation in vendors, but the policy and usability within involved carriers. In countries that are having a top-down security scheme to control their infected IoT devices the infection can be handled fast. In other countries that don’t have such vertical control schemes for IoT yet, they are dwelling variation of horizontal obstacles, from rules and policies, business chain matters and privacy issues , to handle an on-going infection. There is also a factor of unknown vulnerability that has been exploited that can boost infection on a specific region users only.  So, there is no best formula for fast handling, but, it has been proven that the countries that can control IoT devices are less-affected, this is what I think.”

The number of the infection on February 25, 2020 is showing that Taiwan, Hongkong, and Vietnam are having bigger infection ratio compared to other affected countries.

We have learned much experience from the Mirai Fbot re-emerging case. What seems to be the end of a botnet can be a sign of a new beginning, and also by working together we can control IoT malware infection into a very minimum level. If the motivation, bad actors  and the target are there, the same threat could rise again, and again. It is all up to us to police a better policy to prevent it from happening again.

[adrotate banner=”9″]

[adrotate banner=”12″]

Pierluigi Paganini

(SecurityAffairs – Fbot, malware)

[adrotate banner=”5″]

[adrotate banner=”13″]