2021: Information Security Incidents in the Public Domain

22 December 2021

2021 is likely to go down in the history of cybersecurity as the year that has brought at least two game-changing news stories: the successful cyberattack on the Colonial Pipeline company (United States) and the discovery of a critical vulnerability in the Apache Log4j utility, already named the single biggest, most critical vulnerability of the last decade. However, not all industrial cybersecurity incidents, even among the ones that make it to the public domain, receive the proper attention of InfoSec specialists. Today, we’d like to discuss a few incidents that were overshadowed by the bigger news.

At the beginning of 2021, a Russian-speaking IT-community website has published [1] a detailed story of a successful break-in into the corporate network of one of Russia’s largest transportation companies. Unfortunately, this was not the first time when this company has been mentioned in this context. Let’s recount the course of events, draw our conclusions and attempt to come up with a list of practical steps that could make your business less vulnerable in similar situations.

Railroad: Locomotive, 1870

The MikroTik case

What has happened in January 2021? An enthusiast researcher took the initiative to remotely audit the company's technological network, discovering a significant number of various vulnerabilities that allowed easy access to internal resources and services, including CCTV, IP telephony, and tech-process controls. The researcher notified the company but received no feedback, therefore he disclosed the audit results to the general public, hoping to get through to the company management with the aid of the local industrial InfoSec community.

Here’s how the chain of events unfolded. The researcher used the Nmap utility, searching for available proxy servers, assuming that they could exist on the equipment connected to the network while running the default factory settings. One of the scanned public IP addresses turned out to be the interface of the Winbox utility, used for configuring and managing MikroTik RouterOS-based switches. He clearly hit the jackpot, as behind this device, that acted as a proxy-server, another network node was discovered, which had access to a whole segment of the internal corporate network. And this second node did not even have authorization in place.

MikroTik Cloud Router Switch

While trying to identify the network owner, the enthusiast found about a thousand more MikroTik switches and about twenty thousand various IP devices connected to the corporate network, which turned out to be CCTV cameras, IP phones, and workstations. Also, several Intelligent Platform Management Interface consoles that manage server configurations, a few Proxmox VE clusters, several uninterruptible power supply management systems, and other elements typical for a huge IT infrastructure were among the detected nodes. A random analysis of the images pulled from surveillance cameras suggested that the owner was Russia's largest transportation company. Further study of internal services confirmed this guess.

In addition to the general scale of the access gained, the researcher was shocked by a complete absence of active countermeasures on the corporate network. The network segments were not separated by firewalls and a significant amount of equipment used default authorization credentials. One of the conclusions drawn by the researcher is self-sufficient:

As I run an intensive host scan, my connection is not interrupted. Therefore, the company must have never heard of the intrusion detection/prevention systems (IDS/IPS).

The Wi-Fi case

A year before the events described above, on Nov 15, 2019, another enthusiast has published [2] the results of the penetration test conducted on the wireless infrastructure of a high-speed train belonging to the same transportation company, which was carried out during a trip from St. Petersburg to Moscow. A Wi-Fi network, available for all passengers of the train, has used authorization based on a combination of the seat number and the number of the passenger’s ID. This method of authorization led the researcher to a hypothesis that somewhere inside the wireless network segment a repository exists, that contains personal data used for authorization. Using the Nmap and several other public exploits, the enthusiast traversed the network to the server, discovered a significant number of running services, figured out how to deal with the Docker container manager panel cAdvisor, and used root access over SSH to get full access to the files with passengers’ data, related the current and previous train trips.

One of the conclusions drawn by the researcher:

All Wi-Fi users are potentially exposed to traffic sniffing, as their traffic traverses the proxy server on the train. One can easily collect both HTTP and, using special tools, HTTPS traffic.

The community reaction to both cases was so overwhelming, that the company representatives were forced to engage in discussion, partially acknowledging the problems, and announcing that remediation measures were implemented. However, this is not the only possible outcome, and here is another example to complete the picture.

In the summer of 2019, an adversary has gained access to the company's corporate network using compromised employee credentials. He managed to collect a copy of the employees’ and management personal data from an internal storage and obtain an archive of several hundred thousand photos. The acquired data was later made publicly available on one of the dedicated web forums. As a result of the investigation, the perpetrator was identified and prosecuted [3] by the police.

A chain of InfoSec episodes that took place in the time span of two years and concerned the same company. What practical conclusions can be drawn from these episodes? Let’s start with a brief analysis of what happened, identify probable causes, and possible ways to prevent it.

#1. Who’s to blame?

Defining the root cause of events is one of the main quests. Can using SoHo equipment in a production facility explain everything? Or does the problem arise from the integration of multiple products and the blurred responsibility between dozens of contractors?

A national-scale distributed transportation company, with all the resources it manages, is a benchmark example of a complex system. Not just any complex system at that, but a VERY COMPLEX SYSTEM. Efficient operation is impossible without process automation and digitization. Otherwise, every increase in the number of system components would require a nonlinear growth in personnel numbers. And that would make the point of failure shift from technology to a more unpredictable field - the human factor.

We have to give some credit to the aforementioned company: they have been quite successful in automating their business processes. However, automating an extremely complex production results in the creation of enormous IT systems, that inherit the properties of the original EXTREMELY COMPLEX SYSTEM. It’s not the depth of the technological solution stack that poses the challenge, but the sheer number of entities and components comprising the system. When the number of entities is measured in thousands, the manageability of the system and its operation transparency is lost. While the number of different events and control points grows rapidly, the amount of mutual influence possibilities is that number squared.

It should not be assumed that only the architects of the solution, who chose to implement cheaper equipment, or the contractors, who failed proper implementation and commissioning, are at fault. Typical project management mistakes such as cutting project budgets on the go, do not sound like a sufficient explanation either.

Truth is, in recent years, the company has experienced a qualitative leap in system complexity, that exceeded the threshold marking the efficiency of the standard management methods. The operator has lost control over the manageability of the system.

#2: Solving the puzzle

Two possible solutions exist for a situation when system complexity increases rapidly:

1. One option is to perform an administrative breakdown of the entire project into separate sections, delegating control to the dedicated groups of operators. This will automatically lead to an increase in the number of specialists involved (and this increase will be non-linear), which, in the case of a very large system, can in turn boost the company's operational expenses effectively nullifying the initial economic benefit of the automation. This approach is somewhat reminiscent of the «South Park» episode, in which the town's population became so keen on ordering goods from Amazon that as a result, everyone had to work at Amazon's logistics center. Lack of qualified personnel poses another challenge here.

2. Using automation tools to aid IT/InfoSec specialists is a rather viable alternative. These tools can provide visibility into system operation and allow monitoring of the significant parameters for large and geographically distributed systems, while keeping personnel requirements sensible.

As the result, the company may either build a system that would feature a person, responsible for each node or segment operation, and adopt the corresponding staff expenses, or invest in automating the process of collecting, processing, and analyzing the security events generated by the system.

#3. What about the risks?

This example actually features a Critical Information Infrastructure (CII) facility, and in most countries, CII operators are bound to comply with quite a few regulative requirements imposed by the authorities. The fact of public disclosure of the so-called audit results might result in increased attention from various agencies, which may include extraordinary compliance audits and other activities. And it gets worse when consequences are not limited to direct financial losses in form of fines, but also include reputational losses, and, for public companies, ultimately lead to negative capitalization changes.

Let’s review the first incident considering the worst-case scenario. The researcher managed to obtain access to the CCTV system, which means that an adversary could potentially carry out a direct attack on the technical security system, or impact CCTV operation in one or more public places. Station information boards control had also been among the compromised systems, meaning that the intruders might display any sort of messages to an unlimited number of people. These actions may not look hazardous for the primary business processes of the company and general safety, unless we consider them as the first steps of another, possibly more extensive and malicious attack.

#4. Automation options for the workflow of an InfoSec specialist?

The work of an information security officer is not limited to setting up and maintaining asset security at the required level, but also includes a significant amount of paper processes, interaction with internal and external authorities, maintaining operation logs and protocols along with multiple other documents. These clerical tasks require as much labor as initial system design and implementation and are a pain point for many. But what if the bulk of it is switched to automated tools?

The market offers specialized solutions that allow companies to streamline the process of monitoring and controlling the state of security in automated systems. As noted earlier, maintaining the necessary InfoSec state should not rely on a stack of various security products and consoles, effectively making the initial problem worse. As a developer of InfoSec state analysis and monitoring tools for the OT, IoT, and IIoT environments we embrace the approach of providing a single operator console that would encompass all necessary functions for providing a comprehensive view of the protected system. Several key areas can be identified:

Network traffic analysis. 
An in-depth analysis of the network traffic is essential for discovering the assets, determining system topology, and monitoring the interactions that exist between the nodes. Concerning the cases, studied above, this mechanism could have provided the tools necessary to detect the anomalous network activity that had not corresponded with a normal system operation profile, i.e. the connections between the nodes that had been initiated by the enthusiast researcher and should not have existed during regular operation.

Configuration monitoring. 
Collecting configuration files and comparing them with the reference allows precise detection of the unauthorized changes in asset configurations, including hardware, software, and user lists, changes in security settings, services activation/deactivation, and many other areas. In our cases, monitoring the configurations could have enabled the operator to track changes in the assets’ network settings that had been introduced by the researcher in order to gain remote access.

Event analysis.
In most cases, protected assets are able to notify the operators when InfoSec-related events occur. The key here is processing the flow of notifications in a way that allows the important events, that may indicate an ongoing InfoSec incident, to be displayed on the dashboards immediately, while filtering out insignificant events, that may overload the operator or clutter the view. A perfect example here would be a single unsuccessful login attempt versus multiple attempts in a row — the second case is quite likely evidence of a password brute-force attack.

InfoSec compliance control.
Assessing the compliance status of the assets that belong to the protected system with corporate, national, or industry requirements is an essential step to bringing the whole system into compliance. Automating the assessment process relieves the personnel from routine operations allowing human operators to focus on important tasks. In our cases, this assessment could have alerted the operators that a proxy service had been enabled by default on a certain network node.

Vulnerability Analysis. Identification and analysis of known vulnerabilities that exist in the system aids in proper and timely mitigation before those vulnerabilities could be exploited by an adversary. Regarding our case, a vulnerability analysis performed regularly could provide operators with a list of devices running outdated firmware.

Naturally, implementing an InfoSec analysis and monitoring tool implies direct and indirect costs, however, the focus should be on the most significant risk that InfoSec vulnerabilities pose – business processes interruption. Distributed networks and IT systems are built to facilitate the business processes of the company, but how far is an unaddressed vulnerability, that allows external parties to get access to the system with minimum effort, from direct and possibly malicious interference with the core business?

A simple attack example, relevant for logistic operations, would be compromising the cargo tracking database. How would a company assess and clear the chaos in shipments planning and execution? How would it remediate the damage and losses that result from contract terms violation and subsequent claims?

Summary

Finally, let’s pay some attention to the elephant in the room — IT systems, particularly networks, quite often are not just tools for business processes’ support but also facilitate the operation of other systems, that control technological processes. There are known cases when successful ransomware attacks, brought the affected industrial control systems to an inoperable state within a few hours, resulting in downtime for large industrial enterprises, as the technological processes were overly sophisticated for any sort of manual control. Naturally, each unplanned downtime for a large production facility resulted in significant (millions of US dollars) losses for its owner.

The topic of securing the OT infrastructure and the need for companies to comply with all the different rules and regulations for CII owners is nothing new, however reflecting on cases similar to the ones described above proves that it is as relevant and important as ever, especially as digital transformation takes place across multiple industries.

The scope of companies that could benefit from the automation of the InfoSec management processes is not limited to large distributed enterprises but encompasses all businesses that operate complex systems, the security of which are a key factor not only for company success but possibly for the safety of human life.

To sum this up, let’s refer to a famous quote [4] of Bruce Schneier:

Security is a process, not a product.

And remember that processes in both large and complex systems require modern automation tools unless the world is willing to see similar cases, concerning other enterprises.

No matter how experienced and skilled may the employees responsible for system security be, their capabilities are limited by human nature itself: the amount of information they can process, reaction speed, the effects of fatigue, and vision upset along with other factors. Humans have long learned to significantly expand the boundaries of their capabilities through the use of various tools, and it is about time that InfoSec officers obtain and adopt the tools, adequate to the complexity of the tasks they face.

References

[1] https://habr.com/ru/post/536750/
[2] https://habr.com/ru/post/476034/
[3] https://www.interfax.ru/russia/686957
[4] https://www.schneier.com/essays/archives/2000/04/the_process_of_secur.html
[5] https://gudok.ru/newspaper/?ID=1552569

Practical Conclusions

Choosing the right time

A real adversary will choose a proper time for the attack: the moment when the response from the company will be as delayed as possible. Times of peak load on communications networks and power grids, such as the New Year holidays are a perfect example. The nature of the moment camouflages the attack to look at the natural growth of traffic and incident volumes. Forming operational HQs for the entire period associated with peak loads on the systems is highly advisable.

The customer's inertia in accepting the fact of the incident also plays into the adversary’s hands. In the case of the story described in this article, it took the company three weeks until February 10 to advance from the denial phase to normal [5] communications.

Gathering feedback

No matter how talented a security enthusiast may be, often enough they remain introverts and have a hard time communicating the results of their work. Create a simple and easy-to-use channel for communications concerning information security issues, that would have a specialist with appropriate competencies on the other side. It's not exactly about implementing a Bug Bounty program (rewards for external individuals for finding vulnerabilities). First and foremost, it addresses the lack of a mechanism for direct communication between the community and the network security department of the company.

CCTV is a real threat

Don’t be misled by the ease of deploying video surveillance and encapsulating video into IP traffic. Contemporary video surveillance endpoints are full-fledged members of the IoT world, and consequently represent a likely attack target or a potential part of the compromised infrastructure. Typically, there are quite a few cameras in the system, and each has substantial processing power, making them useful in a variety of attack scenarios. Moving the CCTV network (as well as the IP-telephony network) to a separate segment, firewalled from other enterprise networks is very advisable.

Secure the gateways

Special attention must be paid to the points in the infrastructure where internal IT or OT networks and resources interact with public services. These interfaces should be designed and implemented perfectly both in terms of user experience and in terms of InfoSec. UI/UX concerns should never compromise the security of the production network. It’s easy to see how the idea of authorizing the user by using a combination of a seat number and passenger ID number could appeal to the end-users. But in the wrong hands, that same idea could pose a severe problem.

Set up limits and quotas

In addition to granular access control and permission hierarchy, it may be a good idea to set up a quota for the amount of data each account receives. Exceeding this threshold should generate an event in SIEM for further investigation. This allows the operators to identify situations when all of a sudden a user with legal credentials begins to receive unusual amounts of data from the corporate network.

More InfoSec topics from CyberLympha:

Anomalies Detection in Time Series

15 December, 2021

The number of devices, systems, services and platforms belonging to industrial, informational and cyber-physical spheres arоund us increases daily. Usually, we do not... more

Identifying Hosts Class and Detecting Anomalies in Hosts Activity by Passive Network Traffic Monitoring

6 November, 2021

Modern plants, as well as large trains and ships, use data transmission networks. Most of the time this data is quite critical and consequently worth protecting... more

Is Restricted Software Environment a spherical cow?

29 September, 2021

In all fields of science, simplifying the real world in order to successfully develop various theories for the imaginary world is quite normal. Physicists have... more

Contact us

E-mail *
Comment (optional)

68 Circular Road, #02-0168, Singapore 049422
CyberLympha Pte Ltd.