I can remember sometime around late 2001 or early 2002, GREPing Snort logs for that needle in a haystack until I thought I was going to go blind. I further recall around the same time cheering the release of the Analysis Console for Intrusion Databases (ACID) tool which helped to organize the information into something that I could start using to correlate events by way of analysis of traffic patterns.
Skip ahead and the issues we faced while correlating data subtly changed from a one-off analysis to a lack of standardization for the alert formats that were available in the EDR marketplace. Each vendor was producing significant amounts of what was arguably critical information, but unfortunately all in their own proprietary format. This rendered log analysis and information tools constantly behind the 8-ball when trying to ingest all of these critical pieces of disparate event information.
We have since evolved to the point that log file information sharing can be easily facilitated through a number of industry standards, i.e., RFC 6872. Unfortunately, with the advent of the Internet of Things (IoT), we have also created new challenges that must be addressed in order to make the most effective use of data during event correlation. Specifically, how do we quickly correlate and review:
a. Large amounts of data;
b. Data delivered from a number of different resources (IoT);
c. Data which may be trickling in over an extended period of time and,
d. Data segments that, when evaluated separately, will not give insight into the “Big Picture”
How can we now ingest these large amounts of data from disparate devices and rapidly draw conclusions that allow us to make educated decisions during the incident response life cycle? I can envision success coming through the intersection of 4 coordinated activities, all facilitated through event automation:
1. Event filtering – This consists of discarding events that are deemed to be irrelevant by the event correlator. This is also important when we seek to avoid alarm fatigue due to a proliferation of nuisance alarms.
2. Event aggregation – This is a technique where a collection of many similar events (not necessarily identical) are combined into an aggregate that represents the underlying event data.
3. Event Masking – This consists of ignoring events pertaining to systems that are downstream of a failed system.
4. Root cause analysis – This is the last and quite possibly the most complex step of event correlation. Through root cause analysis, we can visualize data juxtapositions to identify similarities or matches between events to detect, determine whether some events can be explained by others, or identify causational factors between security events.
The results of these 4 event activities will promote the identification and correlation of similar cyber security incidents, events and epidemiologies.
According to psychology experts, up to 90% of information is transmitted to the human brain visually. Taking that into consideration, when we are seeking to construct an associational link between large amounts of data we, therefore, must be able to process the information utilizing a visual model. DFLabs IncMan™ provides a feature rich correlation engine that is able to extrapolate information from cyber incidents in order to present the analyst with a contextualized representation of current and historical cyber incident data.
As we can see from the correlation graph above, IncMan has helped simplify and speed up a comprehensive response to identifying the original infection point of entry into the network and then visual representing the network nodes that were subsequently affected, denoted by their associational links.
The ability to ingest large amounts of data and conduct associational link analysis and correlation, while critical, does not have to be overly complicated, provided of course that you have the right tools. If you’re interested in seeing additional capabilities available to simplify your cyber incident response processes, please contact us for a demo at [email protected]
One of my favorite sports, American football, uses a term which has always fascinated me. This term is ‘situational football’ and its whole concept is to react according to the scenario in which you find yourself. American football clubs split their squads into essentially three teams.
–Attack, which is the offensive team and the guys that typically score points.
–Defense, which is the opposite team tasked with stopping the attacking team from scoring points.
–Special teams, which is an often overlooked team. This team can be part of the defense or offense and is typically used for every other play that is not defined as an offensive or defensive setting.
Now, you may be wondering why I am talking about sports in a cyber security blog?!
Well, I always like to relate cyber security industry to other industries and to try to think outside of the box when discussing some of our approaches. That said, I’m going to make a beeline for this idea and start relating this to our thinking:
–Attack, or Red teams, can have a positive impact on your response strategy. Relating your response plans and playbooks directly to common attack methods is advisable and should be used in conjunction with the relevant compliance standards. The actions taken in response to specific attack vectors will usually have a higher success rate than a generic catch-all cyber incident response plans. I would take a lot more comfort knowing I have playbooks designed for a specific threat vector than I would be hoping that one of my generic playbooks would cover it.
–Defense, or Blue Teams, are already a big part of response plans, and ongoing refinement of these plans should coincide with every incident lessons learned. A successful response should still have lessons to consider!
Special Teams are a mix of Red and Blue, of offense and defense. They are best positioned to engage in ‘situational football’ and to enable you to define your approach with more than one mindset, even, in some cases, conflicting mindsets. Using this combined approach will ensure an attackers methodology when searching for enrichment information during incident identification, and the pragmatism of a defender during containment and eradication activities. Having a defined response to each phase of IR is important, but engaging special teams and having the ability to refactor your playbooks on the fly is a key capability when orchestrating an effective cyber security incident response to a dynamic incident.
Unique situations can present themselves at every moment of the game. Our playbook features allow you to make your defense attack-minded by feeding in all the information gathered from your playbooks and allowing you to not be restricted by baseline actions alone. We want your defense to run actions at every point and to allow you to call an audible in any situation that presents itself. The freedom to apply this mindset will drive your incident response teams above and beyond what they see in front of them.
At DFLabs, we not only create playbooks specific to compliance standards and cyber security incident response standards, we also enable you to create and to actively amend your own custom playbooks. Our flexibility ensures that your playbooks can be built on the experience of your Red and Blue teams, in line with adversarial thinking specific to your organization or industry, and to the satisfaction of your corporate, industry and regulatory policies.
Contact us to find out more at [email protected]
“Noise” is a prevalent term in the cyber security industry. DFLabs consistently receives feedback from vendor partners and clients that one of the major issues they face daily is the ability to sift through noise in order to understand and differentiate an actual critical problem from a wild goose chase.
Noise is vast amount of information passed from security products that can have little or no meaning to the person receiving this information. Typically, lots of products are not tuned or adapted for certain environments and therefore would present more information than needed or required.
Noise is a problem to all of us in the security industry, as there are meanings within these messages that are many times simply ignored or passed over for higher priorities. For example, having policies and procedures that are incorrectly identified or adapted or the product is not properly aligned within the network topology.
There is no one security product that can deal with every attack vector that businesses experience today. What’s more disturbing about this paradigm is that the products do not talk to each other natively, yet all these products have intelligence data that can overlay to enrich security and incident response teams.
Cyber incident investigative teams spending a vast number of hours doing simple administration that can be relieved by introducing an effective case management system. Given the sheer volume we can see from SIEM products on a day to day basis we can execute all of the human to machine actions and follow best practice per type of incident and company guidelines through automated playbooks.
Re-thinking about what information is being presented and how we deal with it is the biggest question. There are several ways to manage this:
• Fully automating the noise worthy tasks. If these are consistently coming into your Security Operations Center (SOC) causing you to spend more time on administration than investigation, it may be prudent to schedule the tasks in this manner.
• Semi-Automation of tasks can give your SOC teams more control of how to deal with huge numbers. Automating 95% of the task and then giving this last sign off a manual look over can heavily reduce time if your organisation is against completely automating the process.
• Leverage all your existing products to provide better insight into the incident. For example, leverage an existing active directory to lock out or suspend a user account if they log in outside of normal business hours. Additionally it’s possible to sandbox and snapshot that machine to understand what is happening. A key consideration here is to make sure not to disrupt work at every opportunity. It really is a balancing act, however depending on their privilege you may want to act faster for some users than others.
In 2017, the readiness and capability to respond to a variety of cyber incidents will continue to be at the top of every C-level agenda.
By leveraging the orchestration and automation capabilities afforded by IncMan™, stake holders can provide 360-degree visibility during each stage of the incident response life cycle. This provides not only consistency across investigations for personnel, but encourages the implementation of Supervised Active Intelligence™ across the entire incident response spectrum.
At DFLabs we showcase our capacity to reduce investigative time, incident dwell time all while increasing incident handling consistency and reducing liability. Arming your SOC teams with information prior to the start of their incident investigation will help to drive focus purely on the incidents that need attention rather than the noise.
If you’re interested in seeing how we can work together to grow your incident response capabilities, visit us at https://www.DFLabs.com and schedule a demonstration of how we can utilize what you already have and make it better.