Troubled waters: Cyber-attacks on San Diego and Barcelona’s ports
Last summer’s wave of ransomware attacks compromised port terminals and disrupted global shipping. Since then, cyber security has quickly risen to the top of the agenda for the maritime sector. Earlier this year, another port was hit with ransomware, and then, last week, the ports of Barcelona and San Diego revealed that they had been the victims of further ransomware attacks.
Whilst the 2017 attacks were globally devastating, there was no evidence that they deliberately targeted particular sectors; port terminals were merely caught in the indiscriminate wave of attacks. However, the widespread disruption these attacks caused across industry – from shipping to manufacturing – drew attention to the risk of IT cyber-attacks propagating into the industrial sector’s critical control systems. Operational Technology within industrial environments had previously been kept relatively separate from IT systems, and, consequently, relatively immune from cyber-attack. These attacks showed that the recent trend in integrating and unifying IT and OT systems had now exposed these systems to such indiscriminate attacks.
The increasing convergence of IT and OT systems shows no signs of slowing, however. Hyper-connected ‘smart’ ports are bringing efficiency and precision while cutting costs. Yet, the intertwining of the physical and digital across ports remains a significant challenge for the cyber security teams tasked with their defense. Without rushing to conclusions, it is perhaps no surprise that the Port of Barcelona is in the process of a “Digital Port project,” launched last year to promote the digitization of the port environment.
Although specifics have not yet been revealed, the recent attacks in Barcelona and San Diego appear to be targeted. Perhaps the inadvertent success of last year’s ransomware campaign inspired attackers to pursue the maritime sector specifically. Disruptions to Operational Technology can be highly detrimental to the maritime sector – these systems oversee critical port and ship systems. Any compromise could inflict reputational harm, significant financial losses, and physical damage. That we would see ransomware attacks specifically targeting ports was foreseeable. Many in the industry have been expecting and preparing for such an eventuality over the last 12 months. Now that attackers are actively targeting them, the protection of OT systems has become critical.
Darktrace has deployed AI to a number of companies in the maritime sector to specifically mitigate and defend Operational Technology. These systems are highly customized and bespoke, and therefore unsuitable for the use of off-the-shelf IT solutions. Darktrace’s cyber AI is able to automatically tailor to OT environments and learn a unique sense of ‘self’, regardless of vendor or technology platform.
Our AI is actively defending ports across the world – such as Harwich Haven Authority and Belfast Harbour – and protecting them against both targeted and indiscriminate attacks on their OT and IT systems. Defending these environments requires the ability to protect all technology systems, from the oldest PLCs and SCADA systems, to the newest IoT devices. Whether in the cloud, on a vessel, or on the mainland, Darktrace is able to passively defend your systems and identify cyber-threats in real time, without any impact or disruption.
Healthcare beware: Crypto-mining, malware, and IoT attacks
In the past year, the healthcare industry has been increasingly targeted by advanced cyber-attacks. While a marked rise in medical IoT devices has allowed healthcare companies to become much more efficient, this increase has also opened new avenues for threat actors attempting to infiltrate their networks.
Medical staff now carry multiple connected devices with them, including personal devices that lack appropriate security controls. Confidential patient records and life-critical medical systems run an increased risk of being compromised, and the sensitive nature of the information they contain could impact patient safety and hospital reputation. Financially, healthcare companies are also at greater risk: according to a study by the Ponemon Institute, lost or stolen healthcare records can cost up to 136% more than data breached in other industries.
Going for gold
Towards the end of last year, we observed a noticeable spike in the number of crypto-mining infections within the healthcare sector. In December 2017 alone, the number of crypto-malware attempts on healthcare customers’ systems was 800% higher than in the six months prior and following.
Whilst healthcare companies have always been the target of malware infections, the sudden increase in crypto-malware was significant. This could be attributed to the price of Bitcoin and similar cryptocurrencies skyrocketing around the same time. As their price has now fallen, so have the crypto-mining attempts.
Breaking through Windows
Although 2018 brought with it a decrease in crypto-mining attempts, the healthcare sector experienced an increase in active malware infections captured by sinkhole domains. The infections were varied, with no bias towards botnets, trojans, or ransomware, but were almost entirely united in that the threat actors widely targeted outdated Windows operating systems.
The risks of the EternalBlue SMB vulnerability are now well known. However, as learnt in the aftermath of WannaCry, entire NHS trusts are also susceptible to other unpatched Windows 7 vulnerabilities, including those that facilitate remote code execution and privilege escalation – prime pickings for any malware that successfully enters a system.
Hiding in plain sight
A private medical institution recently trialed Darktrace’s Enterprise Immune System technology through a Proof of Value. Darktrace immediately discovered that an AXIOS spectrometer, a medical IoT device for characterizing materials using x-ray, had been compromised. It had breached hundreds of models, many of a potentially serious nature. The device was continuously making outbound SSH connections to rare external IP addresses, transferring over 1GB of data a week.
Further analysis determined that the compromised medical device was being used to send large volumes of outbound spam mail, resulting in the medical institution’s external IP address being blocked by spam filters. Effectively classified as a sender of junk mail, emails from the medical institution risked falling into recipients’ trash or not being received at all – anything from appointment updates, to the results of cancer scans. Faith in the institution’s ability to handle patient data and uphold its duty of care could have been severely undermined, risking its reputation among prospective patients and service providers.
An AXIOS spectrometer.
Likely C2 beaconing was also noted from this device, indicating that it might have been part of a wider botnet, or network of compromised devices being used to propagate malicious spam malware. On further investigation, at least one of the HTTP connections was to a server utilized within cryptocurrency exchange and bitcoin activity, which suggests a crypto-mining malware presence. The institution’s security team were advised immediately. The device was then isolated, giving the team precious time to conduct further investigation.
What next?
The healthcare sector is a clear target for threat actors, especially considering the wealth of sensitive data such networks safeguard, and the security holes left open in the challenge to continuously maintain and patch highly complex and distributed networks. WannaCry and Petya ransomware were unlikely to have been the last aggressive attacks that successfully exploit such vulnerabilities.
Insider threat is also manifest in healthcare networks. User compliance problems are prevalent, for example, there is a sizable use of Tor as the preferred VPN, widespread use of BitTorrent, and a high volume of illicit uploads to cloud storage services.
Darktrace’s technology has the unique ability to detect and respond to in-progress cyber-attacks that would ordinarily bypass traditional security tools. As threat actors are continually employing novel methods to compromise a network, a growing number of healthcare companies are now having to play catch-up in a fast-evolving threat landscape.
Trusting the cloud: Unencrypted data upload by government body
Managing misconceptions
As the market increasingly moves to the next wave of computing models, over 90% of organizations are expected to adopt hybrid infrastructures by 2020. This move to the cloud brings undeniable benefits for most organizations - from start-ups looking for minimal up-front costs to large organizations striving to boost efficiency, scale on demand, and benefit from constant availability of services and increased agility.
Alongside this growth, the challenge of securing critical data in the cloud has taken on a new dimension. As internal servers are so commonly affected by malware infections or insider threats, there exists a common misconception that the data stored within the cloud is somehow more secure than the data resting on company fileservers. However, this is not necessarily the case – the information stored on cloud infrastructure may be just as (un)safe as any other corporate data store.
Much of this risk comes from the misconception of the network position of cloud servers themselves. Although rented out for use by the company and used every day as part of fundamental business purposes, connections to cloud servers (if not facilitated by a VPN or other strong encrypted channels) cross the perimeter of the network and traverse the public internet. This means that data uploaded to and from the cloud is a prime target for man-in-the-middle attacks, carried out by opportunistic actors hoping to sniff usernames, passwords, and other sensitive details that they could then leverage for direct corporate data theft.
The reality is that while organizations can outsource their IT services, they cannot outsource their security function altogether. In fact, protecting the cloud comes with its own challenges, with most of the existing native security controls and third-party security solutions suffering from significant limitations.
Customer use case
A city government in the United States had outsourced the storage of SQL databases to a cloud storage provider. However, it had not interrogated the protocols that the server by default employed to upload and download information. Addresses, phone numbers, vehicle registration plate numbers: the city government was uploading it all to the external database via unencrypted connections. This highly sensitive data was intended for limited access by select employees within the city government, but the security oversight had made the data available to any attacker clued-up enough to park themselves on the perimeter of the network and collect the data-rich MySQL packets that came their way.
Sample of the plain text containing sensitive data.
Darktrace Cloud detected an unusual SQL connection to a rare external IP from a desktop device within the company. This communication was verified as being SQL-related via packet capture, which then revealed the sensitive public data.
The customer was unaware of this vulnerability, which remained under the radar of its entire security stack. An attacker could easily exploit it to gather material for spear phishing attacks or potentially even identity fraud.
Conclusion
In order to reduce risk and identify atypical or suspicious behavior, full visibility of all cloud services is critical, as hosting data on external servers can create dangerous blind spots and introduce subtle threats that circumvent traditional signature-based tools.
Already over 500 Darktrace customers use Darktrace Cloud to defend cloud environments and SaaS applications, including AWS, Microsoft Azure, Salesforce, and Google Cloud Platform. Darktrace provides businesses with fundamental visibility and real-time threat detection across their entire distributed infrastructures. Through the power of unsupervised machine learning, businesses are now able to confidently tackle the potential risks of data leakage and man-in-the-middle attacks that can affect cloud users.
Beyond the hash: How unsupervised machine learning unlocks the true power of JA3
Introducing JA3
JA3 is a methodology for fingerprinting Transport Layer Security applications. It was first posted on GitHub in June 2017 and is the work of Salesforce researchers John Althouse, Jeff Atkinson, and Josh Atkins. The JA3 TLS/SSL fingerprints created can overlap between applications but are still a great Indicator of Comprise (IoC). Fingerprinting is achieved by creating a hash of 5 decimal fields of the Client Hello message that is sent in the initial stages of an TLS/SSL session.
JA3 is an interesting approach to the increasing usage of encryption in networks. There is also a clear uptick in cyber-attacks using encrypted command and control (C2) channels – such as HTTPS – for malware communication.
The benefits of JA3 for enhancing rules-and-signatures security
These near-unique fingerprints can be used to enhance traditional cyber security approaches such as whitelisting, blacklisting, and searching for IoCs.
Let’s take the following JA3 hash for example: 3e860202fc555b939e83e7a7ab518c38. According to one of the public lists that maps JA3s to applications, this JA3 hash is associated with the ‘hola_svc’ application. This is the infamous Hola VPN solution that is non-compliant in most enterprise networks. On the other hand, the following hash is associated with the popular messenger software Slack: a5aa6e939e4770e3b8ac38ce414fd0d5. Traditional cyber security tools can use these hashes like traditional signatures to search for instances of them in data sets or trying to blacklist malicious ones.
While there is some merit to this approach, it comes with all the known limitations of rules-and-signatures defenses, such as the overlaps in signatures, the inability to detect unknown threats, as well as the added complexity of having to maintain a database of known signatures.
JA3 in Darktrace
Darktrace creates JA3 hashes for every TLS/SSL connection it encounters. This is incredibly powerful in a number of ways. First, the JA3 can add invaluable context to a threat hunt. Second, Darktrace can also be queried to see if a particular JA3 was encountered in the network, thus providing actionable intelligence during incident response if JA3 IoCs are known to the incident responders.
Things become much more interesting once we apply our unsupervised machine learning to JA3: Darktrace’s AI algorithms autonomously detect which JA3s are anomalous for the network as a whole and which JA3s are unusual for specific devices.
It basically tells a cyber security expert: This JA3 (3e860202fc555b939e83e7a7ab518c38) has never been seen in the network before and it is only used by one device. It indicates that an application, which is used by nobody else on the network, is initiating TLS/SSL connections. In our experience, this is most often the case for malware or non-compliant software. At this stage, we are observing anomalous behavior.
Darktrace’s AI combines these IoCs (Unusual Network JA3, Unusual Device JA3, …) with many other weak indicators to detect the earliest signs of an emerging threat, including previously unknown threats, without using rules or hard-coded thresholds.
Catching Red-Teams and domain fronting with JA3
The following is an example where Darktrace detected a Red-Team’s C2 communication by observing anomalous JA3 behavior.
The unsupervised machine learning algorithms identified a desktop device using a JA3 that was 100% unusual for the network connecting to an external domain using a Let’s Encrypt certificate, which, along with self-signed certificates, is often abused by malicious actors. As well as the JA3, the domain was also 100% rare for the network – nobody else visited it:
It turned out that a Red-Team had registered a domain that was very similar to the victim’s legitimate domain: www.companyname[.]com (legitimate domain) vs. www.companyname[.]online (malicious domain). This was intentionally done to avoid suspicion and human analysis. Over a 7-day period in a 2,000-device environment, this was the only time that Darktrace flagged unusual behavior of this kind.
As the C2 traffic was encrypted (therefore no intrusion detection was possible on the payload) and the domain was non-suspicious (no reputation-based blacklisting worked), this C2 had remained undetected by the rest of the security stack.
Combining unsupervised machine learning with JA3 is incredibly powerful for the detection of domain fronting. Domain fronting is a popular technique to circumvent censorship and to hide C2 traffic. While some infrastructure providers take action to prevent domain fronting on their end, it is still prevalent and actively used by attackers.
The only agreed-upon method within wide parts of the cyber-security community to detect domain fronting appears to be TLS/SSL inspection. This usually involved breaking up encrypted communication to inspect the clear-text payloads. While this works, it commonly involves additional infrastructure, network restructuring and comes with privacy issues – especially in the context of GDPR.
Unsupervised machine learning makes the detection of domain fronting without having to break up encrypted traffic possible by combining unusual JA3 detection with other anomalies such as beaconing. A good start for a domain fronting threat hunt? A device beaconing to an anomalous CDN with an unusual JA3 hash.
Conclusion
JA3 is not a silver bullet to pre-empt malware compromise. As a signature-based solution, it shares the same limitations of all other defenses that rely on pre-identified threats or blacklists: having to play a constant game of catch-up with innovative attackers. However, as a novel means of identifying TLS/SSL applications, JA3 hashing can be leveraged as a powerful network behavioral indicator, an additional metric that can flag the use of unauthorized or risky software, or as a means of identifying emerging malware compromises in the initial stages of C2 communication. This is made possible through the power of unsupervised machine learning.
