4 Fundamentals of Good Security Log Monitoring

December Newsletter

By: Randy Franklin Smith

Effective security log monitoring is a very technical challenge that requires a lot of arcane knowledge and it is easy to get lost in the details. Over the years, there are 4 things that stand out to me as fundamentals when it comes to keeping the big picture and meeting the challenge:

1. Just do it
Sometimes organizations prevaricate about implementing a SIEM/log management solution because they aren’t sure they will be able to fully utilize it because of staff or skill shortage and a host of other reasons. Making sure someone is watching your SIEM and follow up on what it finds is certainly important but don’t let that hold you back from implementing SIEM and log management. There are multiple levels on the security monitoring maturity model and you can’t necessarily start off where you’d like to. But you need to be collecting and securely archiving logs no matter what – even if no one is looking at them at all. If you don’t capture logs when they are created you lose them forever; logs will only be there when you need them if you’ve at least been collecting and securely archiving them. That may be the first step on the maturity model but without it you lose accountability and the ability to conduct critical forensics to determine the history and extent of security incidents.

2. Let your environment inform your monitoring
Security analytics technology like SIEMs is getting better and better at recognizing malicious activity out of the box. But there will always be a limit to what shrink wrapped analytics can find. At some point you need to analyze your environment and tailor your monitoring and alerting criteria to take into account what makes your environment different. How is your network divided internally as far as security zones? Which systems should be making outbound connections to the Internet? Which shouldn’t? Which PCs should be making administrative connections to other systems – which shouldn’t. Those are just a few examples. But the more local intelligence you build into your monitoring the better your SIEM will be at recognizing stuff you should investigate.

3. The more secure and clean your environment – the easier it is to detect malicious activity
Here’s a couple examples of what I’m talking about. First a security example: let’s say you lock down your servers so that they can only accept remote desktop connections from a “jump box” you set up which also requires multi-factor authentication. That’s a great step. Now leverage that restriction by teaching your SIEM to alert you when it sees remote desktop connection attempts to those servers from unauthorized systems. APT and other malicious outsiders will likely be unaware of your setup at first and will trip the alarm. Here’s a cleanliness example. Let’s say you have a naming convent for user accounts that allows you to distinguish end user accounts, privileged admin accounts and non-human accounts for services and applications. If you strictly follow that convention you now have all kinds of opportunities to catch bad things as they happen in your environment. For instance if you see a non-human service account trying to logon interactively or via Remote Desktop you may very well have an insider misusing that account or an external attacker who has successfully ran a pass-the-hash attack on that account. Or if you see an administrative account created that doesn’t match naming conventions – that may be a tipoff of an APT creating a back-door account.

4. Leverage a SIEM solution that is intelligent and automates as much as possible
If you are going to follow through on my #2 and #3 recommendations you need a SIEM that frees you up from doing all the obvious stuff that can be packaged by the vendor. EventTracker has been around a long time and has a huge amount of knowledge and support already built into it for many, many different log sources as well as intelligent behavior analysis.
Log monitoring is a rigorous, technical exercise but good SIEM technology frees you up to focus on what makes your environment different and how to leverage those differences to recognize malicious activity as it happens. But if you aren’t at the point to get truly sophisticated in your monitoring don’t let that hold you back from at least collecting and archiving those logs that they are secure and available.