The morning of Friday, May 12 multiple sources in Spain began reporting an outbreak of the ransomware now identified as WannaCry.
Upon learning of these incidents, McAfee immediately began working to analyze samples of the ransomware and develop mitigation guidance and detection updates for its customers.
By Friday afternoon, McAfee’s Global Threat Intelligence system was updated to identify all known WannaCry samples and the company had delivered DAT signature updates to all its customers.
McAfee urges all its customers to ensure these DAT updates have been applied, and furthermore ensure that security updates are applied for all the software solutions they use. For more information, read this Knowledge Center article.
This week’s attacks leveraging the WannaCry ransomware were the first time we’ve seen an attack combine worm tactics along with the business model of ransomware. The weaponization of the Eternal Blue exploit made public weeks ago, and unpatched MS-17-010 Windows OS vulnerabilities by the thousands enabled WannaCry to infect hundreds of thousands of computers, across industries, across continents, and within just a day. Furthermore, these attacks accomplished all this with little or no human involvement, as is typically the case in other ransomware campaigns.
A hybrid of the proven, less the human
WannaCry’s success comes down to its ability to amplify one attack through the vulnerabilities of many machines on the network. The impact of the attack is much greater than what we’ve seen from traditional data ransomware attacks.
Almost all of the ransomware we see in the wild today attack individual users typically through spear-phishing, meaning victims receive an email that appears to be coming from a legitimate source, it lures the victim into clicking on a link or opening an attachment that downloads or executes malicious code on his or her system. But it only impacts that victim’s one computer.
If you think back to the late 90s and early 2000s, when we had Code Red, NIMDA and SQL Slammer, those worms spread really rapidly because they didn’t require a human to take any action in order to activate the malware on the machine. This week’s attacks did something very similar.
We’re still working to determine how a “patient zero” machine became infected, but, once it was, if other machines hadn’t received the MS-17-010 vulnerability patch, they were infected over their network.
Instead of stealing data or damaging other machines, the malware executed a classic ransomware attack, encrypting files and demanding a ransom payment. The attack essentially combined two techniques to produce something that was highly impactful.
With WannaCry, if the configuration of machines within an organization possessed the Microsoft vulnerability (addressed by Microsoft in March), the ransomware could infect one machine and then move very rapidly to spread and impact many other machines that still had not been patched.
What we’ve typically seen with cybercrime is that when any technique is shown to be effective, there are almost always copycats. Given that this appears to have been quite an effective attack, it would be very reasonable for other attackers to look for other opportunities. One of the things that makes that difficult is you need to have a vulnerability in software that has characteristics that enable worm-like behavior.
What’s unique here is that there is a critical vulnerability that Microsoft has patched, and an active exploit that ended up in the public domain, both which created the opportunity and blueprint for the attacker to be able to create this type of malicious ransomware worm capability.
Open for exploit
In the late 90s, it was common practice to leave all sorts of software running on machines even if it wasn’t used. For instance, one of the worms in the 90s took advantage of a vulnerability in a print server which was by default included on all servers even if there wasn’t a printer attached to the configuration of systems. That could enable a worm to connect to that printer port on all of the servers on a network, creating a worm propagation scenario that infected system after system.
A common practice for addressing this since those days is a best practice known as “least privilege,” which allows an application or service to run only the things on a machine or network that that entity needs to complete a task or function specific to its particular role. Least privilege has reduced the chances of the traditional worm scenario, but unpatched vulnerabilities mimmick this “open” element available for exploit, particularly if such vulnerabilities enable things such as file transfer or sharing across systems.
It would be difficult to orchestrate attacks such as the WannaCry campaign, without all the unpatched vulnerabilities, the publicly released exploit, and a set of proven ransomware technologies and tactics at the attacker’s disposal.
To patch or to not to patch
WannaCry should remind IT of the criticality to apply patches quickly. Part of the reason IT organizations hesitate to patch or run an internal quality assurance process is to ensure that there aren’t software incompatibility issues. One way I like to think about this is that whenever a patch must be applied, there is a risk to applying a patch, and a risk to not applying a patch. Part of what IT managers need to understand and assess is what those two risks mean to their organizations.
By delaying deployment of a patch, they can mitigate risk related to application compatibility. By delaying a patch, they are increasing the risk of being compromised by a threat exploiting a vulnerability. IT teams need to understand for each patch, what those levels of risk are, and then make a decision that minimizes risk for an organization.
Events such as WannaCry have the potential to shift the calculus of this analysis. One of the problems we often see in security is that the lack of an attack is sometimes interpreted as having a good defense. Companies that have become lax in applying patches may have not experienced any attacks that take advantage of those vulnerabilities. This can reinforce the behavior that it’s okay to delay patching.
This episode should remind organizations that they really do need an aggressive patching plan in order to mitigate the vulnerabilities in their environment.
Why the hospitals?
Hospitals fall into a category I think of as “soft targets,” meaning hospitals generally focus on patient care as their top priority, as opposed to having the best cyber defenders on staff and best cyber defense technologies in place.
The reason for this is that, traditionally, there was very little incentive for cybercriminals to attack a hospital. They could potentially steal patient records or other data, but the total value of data from a hospital would typically be less than that of the bulk data stolen from other industries such as financial services.
What ransomware has done as a criminal business model is provide an incentive to attack any organization. Given that criminals are demanding a ransom, it’s far easier to exploit an organization with weaker cyber defenses than an organization with stronger cyber defenses, which is why we’ve seen hospitals, schools, municipal police departments, and universities become victims of ransomware over the last year. While we’re now starting to see the targeting of “harder” organizations as well, at least for now, there are a lot of opportunities for criminals to continue to target these soft target organizations.
Although this attack is something new, and something we need to be thoughtful of, when we see such a vulnerability occur in the wild, and an exploit published that could be used by cybercriminals, we should always expect and be prepared for this kind of attack, and many more copy-cat attacks following soon after.