Mirai has been making a zombie army of swaths of the internet of things, so it is no wonder that manufacturers are taking steps to protect against it. However, one IoT device manufacturer’s failed attempt to protect its products against the botnet (used in massive DDoS attacks) shows how challenging this climate has become. When IoT-maker XiongMai, based in China, attempted to patch its devices so that the malware would be blocked, the result was described as a “terrible job” by security consultant Tony Gee.
Gee explained that he took products from the manufacturer to a trade convention, the Infosecurity Europe Show. The Floureon digital video recorders (DVRs) used in Gee’s demo did not have telnet open on port TCP/23 – but shutting off telnet access was insufficient as a defense.
Gee went through port 9527 via ncat. The passwords matched those of the web interface, and it was possible to open a command shell. Within the command shell, Gee opened a Linux shell and established root access. From the root user position, it was simple to enable telnet.
For devices that have telnet closed down, the device is hackable via shell and restarting the telnet daemon, explained Gee, adding ominously, “And we have Mirai all over again.”
- Tale of an immortal zombie
- How could Mirai grow larger?
- The doom and gloom of Mirai
- How to protect yourself from DDoS
- Layers of protections
- What this all means “on the ground”
Tale of an immortal zombie
Mirai is changing, much to the frustration of those who care about online security. Prior to this point, malware that was infecting IoT devices (such as routers, thermostats, and CCTV cameras) could be cleared away with a reboot.
A method was discovered in June that could be used to remotely access and repair devices that have been enslaved by the botnet. The flip side of this seemingly good news is that the same avenue is a way that a Mirai master can generate reinfection post-reboot – so researchers did not release details.
Notably, BrickerBot and Hajime also have strategies that try to create a persistent, “immortal” botnet.
The researchers did not provide any specific information about the vulnerability out of concern that it would be used by a malicious party. The firm did list numerous other weaknesses that could be exploited by those using the botnet.
How could Mirai grow larger?
What are other possible paths of exploit that would allow Mirai to grow even larger than it is now? Those include:
- DVR default usernames and passwords that can be incorporated into the worm element of Mirai, which uses brute-force methods through the telnet port (via a list of default administrative login details) to infiltrate new devices.
- Port 12323, an alternative port used as telnet by some DVR makers in place of the standard one (port 23).
- Remote shell access, through port 9527, to some manufacturer’s devices through the username “admin” and passwords “[blank]” and “123456.”
- One DVR company that had passwords that changed every single day (awesome), but published all the passwords within its manual on its site (not awesome).
- A bug that could be accessed through the device’s web server, accessible through port 80. This firmware-residing buffer overflow bug currently exists in 1 million web-connected DVR devices.
- Another bug makes it possible to get password hashes from a remote device, using an HTTP exploit called directory traversal.
The doom and gloom of Mirai
The astronomical expansion of Mirai is, at the very least, disconcerting. One recent report highlighted in TechRepublic found that Internet of Things attacks grew 280% during the first six months of 2017. The botnet itself is at approximately 300,000 devices, according to numbers from Embedded Computing Design. That’s the thing: Mirai is not fundamentally about IoT devices being vulnerable to infection. It’s about the result of that infection – the massive DDoS attacks that can be launched against any target.
Let’s get back to that infected and unwitting Frankenstein-ish army of “things” behind the attacks, though – it could grow through changes to the source code by hackers, updating it to include more root login defaults.
The botnet could also become more sophisticated and powerful as malicious parties continue to transform the original so that it has more complex capacities to use in its DDoS efforts. Today it has about 10 vectors or modes of attack when it barrages a target, but other strategies could be added.
How to protect yourself from DDoS
Distributed denial of service attacks from Mirai really are massive. They can push just about any firm off the Internet. Plus, the concern is not just about that single event of being hammered by false requests. Hackers first open up with a toned-down attack, a warning shot that is often not recognized as a pre-DDoS by custom in-house or legacy DDoS mitigation tools (as opposed to a dedicated DDoS mitigation service). These dress-rehearsal attacks, usually measuring under 1 Gbps and lasting 5 minutes or less, are under the radar of many DDoS protection solutions that have settings requiring attack traffic to be more substantial.
When DDoS started more than 20 years ago, engineers would use a null route, or remote trigger blackhole, to push the traffic away from the network and prevent collateral damage to other possible victims.
Next, DDoS mitigation became more sophisticated: traffic identified as problematic on a network was redirected to a DDoS scrubbing service – in which human operators analyzed attack traffic. This process was inefficient and costly. Also, remediation often did not get started right away following detection.
Now, DDoS protection both must be able to “see” a DDoS developing in real-time and have the ability to gauge the DDoS climate for trends, generating proactive steps to mitigate an attack. Enterprise-grade automatic mitigation protects you from these attacks and maintains your reliability.
Layers of protections
There are various levels at which distributed denial of service can be and should be challenged and stopped. First, a DDoS protection service against real and present threats, built by a strong provider, can effectively keep you safe from these attacks – but there are other efforts that can be made as well. Internet service providers (ISPs) can also protect the web by removing attack traffic before it heads back downstream.
Defense should really be at all levels, though. The people who make the pieces of the IoT – software, firmware, and device manufacturers – should build it with protections in place from the start. Installers and system admins should update passwords from the defaults and patch any intrusions as possible.
What this all means “on the ground”
It’s important to recognize that this issue is not just about security firms, device manufacturers, and criminals. It’s also about, really, all of us: the home users of devices, such as the DVR. (If you don’t know, a DVR is a device that records video on a mass storage device such as an SD memory card or USB flash drive… when it isn’t busy being used in botnet attacks).
The home user should be given reasonable security advice. Many users do not respond quickly when new patches are released. IoT devices are often built just strongly enough that they can operate; security is not a priority. That is bad – but that means users need to do their homework on security prior to purchase. They need to change the password from default to complex and randomized ones.
Protect yourself from Mirai
What can you do to keep your business safe from Mirai and other DDoS attacks? At Total Server Solutions, our DDoS mitigation & protection solutions keep your site up and running, your content flowing, and your customers buying, seamlessly. How does it work?