Host Intrusion Prevention Systems: Defense-in-Depth's Best Friend
Viruses, worms, spyware, malware, adware, X-ware...but where? Try everywhere! Todays biggest security issues are highly complex and take many forms in their efforts to penetrate corporate and home defenses. It seems that just about everyday, security professionals hear about the latest "something-ware" they will need to combat in their environment. A defense plan, while often a layered approach, hinges most successfully on the security mechanisms located on the endpoint itself. Let us take a look at these common issues and how host Intrusion Prevention Systems (IPS) can become the most crucial piece of your corporate layered security plan.
Defense-in-depth is the phrase most commonly used to describe the many-layered components securing computing environments. The components vary in their capabilities and where they are located. Many of these security mechanisms are familiar to most of us, such as firewalls, router ACL's, network Intrusion Detection/Prevention Systems (IDS/IPS), content filtering, and host IPSs. By placing these components, as appropriate, throughout your network, you can ensure any attack must pass through one or more of these defenses. This gives you more chances to identify an attack and provide multiple locations to enforce restrictive actions on the attack flows, thereby limiting your exposure to the attacks propagation. Placing security controls throughout your environment is an extremely successful strategy and should be the norm in todays network architectures. The problem with this architecture, however, is that the security mechanisms used as the enforcement and detection points often have limited capabilities.
The majority of network security products' detection capabilities are static. For example, if you look at the security provided by a network ACL in a router or a policy rule in a firewall, you can see that these devices are limited to detecting and enforcing a very specific rule set based on the traffic they detect. Typically, this means only interpreting the Layer 3 and Layer 4 IP headers, comparing these to the internal rules set and the current connection table, and deciding whether the traffic is allowed. This does not take into account what is actually being sent in those communication flows; it only allows these devices to decide who can send and receive information based on addressing information.
A more advanced form of network inspection is found in today's newest network security technology, such as Cisco ISR routers, network IPS, and firewalls with deep-packet inspection capabilities. These devices can dig into the network flows to inspect the contents of the transmissions. This capability allows security professionals to feel more assured about what is actually being transmitted across their networks. Implementing a combination of legacy and next-generation network controls does give network administrators more confidence about their security implementation, but this architecture is still flawed because it relies on four facts to be true for its success:
- The endpoint being secured must be connected to the network.
- The traffic flows (viral, worm, or X-ware) must flow through the network inspection devices.
- The traffic must be unencrypted.
- There must be knowledge of the attack in the form of a signature or well-known attack transmission pattern.
Today's portable laptops, which are often taken outside a protected network, easily remove facts one and two from the equation. This is not due to intentional, premeditated circumvention of security controls but a result of regular business practices. Additionally, the third fact will often be removed from the equation simply by using common encryption protocols; SSL tunneling is a basic example of this circumvention. In many cases, this is also not a premeditated act of corporate defiance, rather users using SSL to transmit and receive personal e-mail from their web-based free-mail service (complete with viral attachments). Finally, the fourth fact is often the most misunderstood. Network security systems, such as IDS and IPS, that rely upon signatures or patterns to accurately identify a virus or other attack are only as accurate as the currently running definition file. Day-zero, or previously unseen attacks, will still slip through undetected until a pattern file update occurs. You can see how even the best layered security deployment can be flawed in its implementation. So how does one solve this architectures shortcomings? You implement host Intrusion Prevention Systems (IPS).
Host IPS is security software that is loaded on each PC and server you want to protect. This software monitors and tracks how processes running on each system interact with each other and the operating system itself. This software control mechanism has policy-based rules configured by the security staff that control how applications are allowed to run. These rules vary in their functions but can control how processes access the network, disk drives, memory, COM components, registry, and various programming APIs, among other items. This allows you to have extremely granular control over how systems and their software function. By implementing host IPS, you are not only protecting your systems from unauthorized access and X-ware Trojan horse-like behavior (including day-zero attacks) but also implementing a software package capable of enforcing your corporate security policy and acceptable usage documents.
So how does host IPS fare against our list of defense-in-depth shortcomings? First, with host IPS loaded on your systems, the PC is no longer required to remain connected to your network. Host IPS software continues to protect the endpoint regardless of where it is connected. Second, since we are concerned with the X-ware, viral, or worm interaction on our system, we no longer care what the flow over the network looks like. Third, since the PC itself is the termination point for any encrypted communication, the host IPS will see the intended negative interaction as clear transmissions instead of what may have been witnessed as network interrogation points. Finally, host IPS such as the Cisco Security Agent does not rely upon signature updates to be successful at preventing attacks. Cisco Security Agent relies on its knowledge of preventing attack behaviors and actions rather than attempting to detect known patterns.
As you can see, the addition of a host IPS solution to your current security architecture can vastly improve your chances of truly securing your environment and ensure your nights and weekends are spent at home with your family rather than in the office cleaning infected systems.