The WannaCry ransomware attack has left security teams around the world scrambling to make sure they are protected and to assess whether they have been victimized. To protect themselves, organizations need to have visibility into which systems are vulnerable and be able to rapidly roll out patches. To understand whether they have been targeted, they need visibility into the channels over which the ransomware is distributed. To understand whether they have been infected, they need visibility into the endpoints.
Over the past weekend, I was bombarded with questions from current customers, potential customers, former colleagues, friends, and family. Am I vulnerable? How do I protect myself? How do I know if I’ve been hit? What do I do if I’ve been hit? What can you do to help me?
This post focuses primarily on the last question. What can we at Sift Security do to help an organization respond to a massive ransomware attack? I break this down into four categories, visibility, analytics, intelligence, and response.
What are the first questions that come to mind whenever information about a new threat is published? Am I vulnerable? Am I being targeted? Have I been hit?
Time is the enemy. You need to answer these questions immediately. Your employees are nervous, your management team is breathing down your neck. You need visibility. You need to have information about all of the devices on your network readily available, and you have to have the right information available to answer the relevant questions.
The first thing that Sift Security does is help provide this visibility. We centralize and display all of the relevant context on a single pane of glass. What’s more, we have put our effort into making a tool that is visual and easy to understand. We make it possible for a user to see what hosts are vulnerable, what users are on those hosts, whether there has been lateral movement from any of those hosts. As an example, the following screenshot shows a user, their workstation, and its IP. The red edges are lateral movement from the workstation to other hosts on the network.
No organization strives to be reactionary. No one wants to wait until a threat has been discovered by someone else to be able to detect it. Nobody wants to wait for IOCs to be written, disseminated, ingested, and cross-referenced to know if they are a victim. Everyone strives to be proactive. We all want to say we did everything we could to be prepared, we saw it happening when it happened, and we responded immediately to limit the damage.
Now that WannaCry is all over the news, we know exactly how to detect it: Grab the latest IOCs and cross reference them against our data. But what if you were hit before it was all over the news? Before we had IOCs? Before we know which vulnerability it exploited? Before we had a name for it?
The answer: data analytics. There are many things that WannaCry does that are similar to many other malware samples and are detectable with the right analytics. One example of how WannaCry is detectable and how analytics can detect it follows.
WananCry spreads like a worm, relying heavily on SMB to infect new machines. This is a particularly effective technique for spreading within an organization, where files and printers are shared over SMB. An infected host immediately starts trying to connect to all SMB shares on the same subnet and across the internet. This behavior can be automatically detected through analytics, by having observed normal SMB activity in the past, and alerting on this new, unusual activity. There are many ways in which this activity is potentially unusual:
- Hosts connect to SMB shares that they usually don’t connect to.
- Hosts that don’t normally use SMB start using it.
- Clients see a spike in SMB activity.
- SMB shares themselves see a spike in activity (valid and invalid authentications)
As important as each of the individual descriptions of unusual activity above is the correlation of those descriptions. They are observable on clients, on SMB servers, and on the network. Observing these behaviors on multiple hosts and at multiple levels is more significant than observing isolated events.
At Sift Security, our analytics engine is constantly looking for these and other abnormal behaviors. We are also constantly looking for correlated patterns of abnormal behaviors. Using our graph analytics, we can present clusters of hosts within your network that are being targeted.
In the previous section, I dismissed traditional IOCs and focused on analytics. But IOCs and threat intel feeds are incredibly important. Once they are known, they provide a means of rapidly assessing whether you are being victimized by a specific threat. Combined with analytics, intelligence can provide a very accurate assessment of whether an organization has been targeted and the scope of the attack. We combine threat intelligence alerts together with third party alerts and analytics to make sure we present our users with the most complete picture of an attack.
Rapid response is key, especially with a threat like WannaCry that is spreading like a worm. For this reason, Sift Security has a response component that integrates with third party products. For example, if we determine a host to be compromised, we can, within the same pane of glass, right click a host and isolate it on the network. Maybe we also want to isolate the hosts that we saw it connect to as well. We’ve seen cases where spear phishing emails were used as the initial attack vector. Deleting these emails and blocking these senders is key to limiting exposure of users.