A lot of our research is conceptual, so we like to wrap up with a scenario. This helps make the ideas a bit more tangible, and provides context for you to apply it to your particular situation. To illuminate how the Security Analytics Team of Rivals can work, let’s consider a scenario involving a high-growth retailer who needs to maintain security while scaling operations which are stressed by that growth.
So far our company, which we’ll call GrowthCo, has made technology a key competitive lever, especially around retail operations, to keep things lean and efficient. As scaling issues become more serious they realize their attack surface is growing, and may force shortcuts which exposure critical data. They has always invested heavily in technology, but less in people. So their staff is small, especially in security.
In terms of security monitoring technologies in place, GrowthCo has had a SIEM for years (thanks, PCI-DSS!). They have very limited use cases in production, due to resource constraints. They do the minimum required to meet compliance requirements.
To address staffing limitations, and the difficulty of finding qualified security professionals, they decided to co-source the SIEM with an MSSP a few quarters ago. The MSSP was to help expand use cases and take over first and second tier response. Unfortunately the co-sourcing relationship didn’t completely work out. GrowthCo doesn’t have the resources to manage the MSSP, who isn’t as self-sufficient as they portrayed themselves during the sales process. Sound familiar?
The internal team has some concerns about their ability to get the SIEM to detect the attacks a high-profile retailer sees, so they also deployed a security analytics product for internal use. Their initial use case focused on advanced detection, but they want to add UBA (User Behavior Analysis) and insider threat use cases quickly.
The challenge facing GrowthCo is to get its Team of Rivals – which includes the existing SIEM, the new security analytics product, the internal team, and the co-sourcing MSSP – all on the same page and pulling together on the same issues. Let’s consider a few typical use cases to see how this can work.
Detecting Advanced Attacks
GrowthCo’s first use case, detecting advanced attacks, kicks off when their security analytics product an alert. The alert points to an employee making uncharacteristic requests on internal IT resources. The internal team does a quick validation and determines that it seems legitimate. That user shouldn’t be probing the internal network, and their traffic has historically been restricted to a small set of (different) internal servers and a few SaaS applications.
To better understand the situation, context from the SIEM can provide some insight into what the adversary is doing across the environment, and support further analysis into activity on devices and networks. This is a different approach to interacting with their service provider. Normally the MSSP gets the alert directly, has no idea what to do with it, and then sends it along to GrowthCo’s internal team to figure out. Alas, that typical interaction doesn’t reduce internal resource demand as intended.
But giving the MSSP discrete assignments like this enables them to focus on what they are capable of, while saving the internal team a lot of time assembling context and supporting information for eventual incident response. Returning to our scenario: this time the MSSP identifies a number of privilege escalations, configuration changes, and activity on other devices. Their report details how the adversary gained presence and then moved internally, to compromise the device which ultimately triggered the SIEM alert.
This scenario could just as easily have started with an alert from the SIEM, sent over from the MSSP (hopefully with some context) and then used as the basis for triage and deeper analysis, using the security analytics platform. The point is not to be territorial about where each alert comes from, but to use the available tools as effectively as possible.
Hunting for Insiders
Our next use case involves looking for potentially malicious activity by employees. This situation blurs the line between User Behavioral Analysis and Insider Threat Detection, which share technology and process. The security analytics product first associates devices in use with specific users, and then gathers device telemetry to provide a baseline of normal activity for each user. By comparing against baselines, the internal team can look for uncharacteristic (anomalous) activity across devices for each employee. If they find something the team can drill into user activity or pivot into the SIEM and use the broader data it aggregates to search and drill down into devices and system logs for more evidence of attacker activity.
This kind of analysis tends to be hard on a SIEM, because the SIEM data model is keyed to devices, and SIEM wasn’t designed to performa a single analysis across multiple devices. That does not mean it is impossible, or that the SIEM vendors aren’t adding more flexible analysis, but SIEM tends to excel when rules can be defined in advance to correlate. This is an example of choosing the right tool for the right job. A SIEM can be very effective in mining aggregated security data when you know what to look for.
Finally, you can also use the Team of Rivals to deal with the other class of ‘adversary’: an auditor. Instead of having an internal team spend a great deal of time mining security data and formatting reports, you could have an MSSP prepare initial reports using data collected in the SIEM, and have the internal team do some quick Q/A, optimizing your scarce security resources. Of course the service provider lacks the context of the internal team, but they can start with the deficiencies identified in the last audit, using SIEM reports to substantiate improvements.
Once again, being a little creative and intelligently leveraging the various strengths of the extended security team, a particularly miserable effort such as compliance reporting can be alleviated by having the service provider do the heavy lifting, relieving load on the internal team. But the internal team bears ultimate responsibility, and so must still ensure the completeness and quality of reports, providing additional documentation as necessary.
As we have said through this series, it is no longer an either/or decision between security analytics and SIEM. Or even insourcing versus outsourcing security monitoring. You need to establish a team who can do what they do best. It is always frustrating to use a technology for something it’s not built to do, and just as frustrating to expect a service provider to do things beyond their capability – whatever they said during the sales cycle. So don’t do that, but instead build your monitoring program to give all parties the best chance of success.
We have showed these technologies aren’t mutually exclusive, and how a team can leverage them both, across a variety of internal and external teams and tools, to solve key problems. The goal is not, of course, to run multiple overlapping technologies in parallel forever. We expect you to eventually be able to move to a single strategic security monitoring platform. But not today, so you’ll need to wrangle a Security Analytics Team of Rivals in the meantime.