It starts right there in PCI-DSS Requirement 1. Install and maintain a firewall configuration to protect cardholder data. Since it’s the first requirement, firewalls must be important, right? Not that PCI is the be all, end all of security goodness, but it does represent the low bar of controls you should have in place to defend against attackers. As the line of first defense on a network, it’s the firewall’s job to enforce a set of access policies that dictate what traffic should be allowed to pass. It’s basically the traffic cop on your network, as well as acting as a segmentation point between separate networks.

Given the compliance mandates and the fact that firewalls have been around for over 20 years, they are a mature technology which every company has installed. It may be called an access router or UTM, but it provides firewall functionality. The firewalls run on a set of rules that basically define what ports, protocols, networks, users, and increasingly applications, can do on your network. And just like a closet in your house, if you don’t spend time sorting through old stuff it can become a disorganized mess, with a bunch of things you haven’t used in years and don’t need any more. That metaphor fits your firewall rule base – when we talk to security administrators and they admit (often in a whisper) to having thousands of firewall rules, many of which haven’t been looked at in years.

The problem is that, like your closet, this issue just gets worse if you put off addressing the issue. And it’s not like rule bases are static. You have new requests coming in to open this port or allow that group of users to do something new or different pretty much every day. The situation can get out of control quickly, especially as you increase the number of devices in use. That creates significant operational and security problems, including:

1. Attack Surface Impact: When a change request comes in, how many administrators actually do some work to figure out whether the change would create any additional attack surface or contradict existing rules? Probably not enough, so firewall management – first and foremost – must maintain the integrity of the protection the firewall provides.
2. Performance Impact: Every extra rule in the rule base means the firewall may need to do another check on every packet that comes through, so more rules impact device the performance. Keep in mind that the order of your rule set also matters, as the sooner you can block a packet, the less rules you will have to check, so the rules should be structured to eliminate connections as quickly as possible.
3. Verification: If a change was made, was it made correctly? Even if the change is legitimate and your operational team is good there will still be human errors. So another problem with firewall management at scale is verifying each change.
4. Weak Workflow and Nonexistent Authorization: What happens when you receive a rule change? Do you have a way to ensure each change request is legit? Or do you do everything via 10 year old forms and/or email? Do you have an audit trail to track who asked for the change and why? Can you generate documentation to show why each change was made? If you can’t it is probably an issue, because your auditor is going to need to see substantiation.
5. Scale: The complexity of managing any operational device increases exponentially with every new device you add. Firewalls are no exception. If you have a dozen or more devices, odds are you have an unwieldy situation with inconsistent rules creating security exposure.
6. Heterogeneity: Many enterprises use multiple firewall vendors, which makes it even more difficult to enforce consistent rules across a variety of devices.

As with almost everything else in technology, innovation adds a ton of new capabilities but increases operational challenges. The new shiny object in the firewall space is the Next-Generation Firewall (NGFW). At a very high level, NGFWs add the capability to define and enforce policies at the *application( layer. That means you can finally build a rule more granular than ALLOW port 80 traffic – instead defining which specific web-based applications are permitted. Depending on the application you can also restrict specific behaviors within an application. For example you might allow use of Facebook walls but block Facebook chat. You can enforce polices for users and groups, as well as certain content rules (we call this DLP Lite). The NGFW is definitely not your grand-pappy’s firewall, which means they dramatically complicate firewall policy management.

But network security is going through a period of consolidation. Traditionally separate functions such as IPS and web filtering are making their way onto a consolidated platform that we call the Perimeter Security Gateway (PSG). Yup, add more functions to the device and you increase policy complexity – making it all the more important to maintain solid operational discipline when managing these devices. In any sizable organization the PSG rule base will be difficult to manage manually. Automation is critical to improving speed, accuracy, and effectiveness of these devices.

We are happy to get back to our network security roots and documenting our research on the essentials of managing firewalls. This research is relevant both to classical firewalls and PSGs. In Firewall Management Essentials we will cover the major areas of managing your installed base of firewalls. Here is our preliminary plan for posts in the series:

1. Automating Change: Firewall management comes down to effectively managing exceptions in a way that provides proper authorization for each change, evaluating each one to ensure security is maintained (including ensuring new attack paths are not introduced), auditing all changes, and rolling back in the event of a problem.
2. Optimizing Performance: Once the change management process is in place, the next step is to keep the rule set secure and optimized. We will discuss defining the network topology and identifying ingress and egress points to help prioritize rule sets, and point out potential weaknesses in security posture.
3. Managing Access: With a strong change control process and optimized rule set, the last aspect of comprehensive firewall management is to ensure that any and all attack paths are understood and managed effectively.
4. Quick Wins: We will wrap up this series with deployment architectures and integration with existing systems (such as attack path analysis, vulnerability management, help desks, reporting & GRC, etc.). We will also work through a deployment scenario so you understand how these tools can fit into your environment.

Don’t the Firewall Vendors Do This?

We should probably tackle the elephant in the room before we jump in. You are probably asking why the firewall vendors don’t do this. That’s a good question and one we have pondered for years. There wouldn’t be a market for firewall management tools if firewall vendors did their jobs better. But they don’t, so a market developed. When you take a step back, the firewall vendors don’t have very good tools to optimize their firewall rule bases. The cynic (who could that be?) would say that firewall vendors don’t want to optimize their devices because a poorly configured rule base kills device performance, and might force a premature upgrades. Yeah, that’s pretty cynical.

The firewall vendors also don’t have very good workflow or compliance reporting as part of their management capabilities. Finally, there isn’t any kind of real incentive for vendors to support heterogenous firewalls, so if you have multiple vendors installed then vendors aren’t going to be much help. And it’s not clear that any of the incumbents are going to get religion about any of these topics any time soon, so you need to look at third party tools if you want to improve management of your firewalls. So we will tell you what you need to know with this series.

As with all our blog series, the research will use our Totally Transparent Research methodology. And we would like to thank the folks at Firemon for potentially licensing the content at the end of this series. With that, we will jump into firewall change management in our next post.