When hearing the term firewall, most people think of a product that inspects network traffic at the network and transport layers of the OSI Reference Model and makes pass or filter decisions. In terms of products, dozens of firewall types exist. They are differentiated by the data source they inspect (e.g., network traffic, host processes, or system calls) and the depth to which they inspect those sources. Almost any device that inspects communication and decides whether to pass or filter it could be considered a firewall product.

Marcus Ranum, inventor of the proxy firewall and the implementer of the first commercial firewall product, offered a definition of the term firewall in the mid-1990s when he said, "A firewall is the implementation of your Internet security policy." 1 This is an excellent definition because it is product-neutral, timeless, and realistic. It applies equally well to the original firewall book, Firewalls and Internet Security by William R. Cheswick and Steven M. Bellovin (Addison-Wesley Professional, 1994), as it does to the book you're reading now.

1 Computer Security Journal, Vol. XI, No. 1, Spring 1995 ( hownot.htm)

In the spirit of Ranum's definition, a firewall could also be considered a policy enforcement system. Devices that inspect and then pass or filter network traffic could be called network policy enforcement systems. Devices that inspect and then pass or filter host-centric activities could be called host policy enforcement systems. In either case, emphasis on policy enforcement focuses attention on the proper role of the firewall as a device that implements policy instead of one that just "stops bad stuff."

With respect to "bad stuff," it's reasonable to ask if firewalls even matter in today's enterprise. Properly configured traditional network firewall products basically deny all but allowed Internet protocols, IP addresses, TCP/UDP ports, and ICMP types and codes. In the modern attack environment, this sort of defense is entirely insufficient. Restricting those exploitation channels is necessary to restrict the ingress and egress paths to a target, but network and transport layer filtering has been a completely inadequate counter-measure for at least a decade.

In 2007, the most effective way to compromise a client is to entice the user to activate a malicious executable, send the user a link that hosts malicious content, or attack another client-side component of the user's computing experience. In many cases, exploitation doesn't rely on a vulnerability that could be patched or a configuration that could be tightened. Rather, attackers exploit weaknesses in rich-media platforms like JavaScript and Flash, which are increasingly required for browsing the Web today.

In 2007, the most effective way to compromise a server is to avoid the operating system and exploit the application. Web applications dominate the server landscape, and they are more likely to suffer from architectural and design flaws than from vulnerabilities that can be patched. In the late 1990s, it was fashionable to change the prices for the items in one's shopping cart to demonstrate insecure web applications. Thanks to Ajax, almost a decade later the shopping cart is running on the client and users are again changing prices—and worse.

All of this makes the picture seem fairly bleak for firewall products. Many have adapted by incorporating deep packet inspection or operating at or beyond the application layer of the OSI Reference Model. Others operate as intrusion prevention systems, using a clever marketing term to differentiate themselves in a seemingly commoditized market. Is there a role for firewalls, especially open source products, in the age of client-side attacks and web application exploitation?

The answer is yes—and you are reading one approach right now. Michael Rash is a pioneer in the creative use of network technologies for defensive purposes. The security research and development world tends to be dominated by offensive tools and techniques, as a quick glance at the speakers list for a certain Las Vegas hacker convention will demonstrate. Bucking this trend, Michael continues to invent and improve upon methods for protecting assets from attack. After getting a look at the dark side at an offensive conference, almost all of us return to the seemingly mundane job of protecting our enterprises. Thanks to this book, we have an additional suite of programs and methods to make our jobs easier.

While reading a draft of this book, I identified a few themes. First, host-centric defense is increasingly important as devices become self-reliant and are exposed to the Internet. An extreme example of this evolution is the introduction of IPv6, which when deployed as intended by its progenitors restores the "end-to-end" nature of the original Internet. Of course, end-to-end can be translated into attacker-to-victim, so additional ways for hosts to protect themselves are appreciated. Linux Firewalls will teach you how hosts can protect themselves using host-based firewalls and tools.

Second, despite the fact that hosts must increasingly defend themselves, host-centric measures alone are inadequate. Once a host has been compromised, it can no longer be responsible for its own defenses. Upon breaching a system, intruders routinely disable host firewalls, antivirus software, and other protective agents. Therefore, network-centric filtering devices are still required wherever possible. An endpoint controlled by a victim can only use the communication channels allowed by the network firewall, at least limiting the freedom to maneuver enjoyed by the intruder. Linux Firewalls will also teach you how network devices can protect hosts.

Third, we must look at creative ways to defend our assets and understand the attack landscape. Single Packet Authorization is a giant step beyond port knocking if one wants to limit access to sensitive services. Visualization helps render logs and traffic in a way that enables analysts to detect subtle events of interest. After reading this book, you may find additional ways to leverage your defensive infrastructure not anticipated by others, including the author.

I'd like to conclude these thoughts by speaking as a book reviewer and author. Between 2000 and mid-2007, I've read and reviewed nearly 250 technical books. I've also written several books, so I believe I can recognize a great book when I see it. Linux Firewalls is a great book. I'm a FreeBSD user, but Linux Firewalls is good enough to make me consider using Linux in certain circumstances! Mike's book is exceptionally clear, organized, concise, and actionable. You should be able to read it and implement everything you find by following his examples. You will not only familiarize yourself with tools and learn to use techniques, but you will be able to appreciate the author's keen defensive insights.

The majority of the world's digital security professionals focus on defense, leaving offense to the bad guys, police, and military. I welcome books like Linux Firewalls that bring real defensive tools and techniques to the masses in a form that can be digested and deployed for minimum cost and effort. Good luck—we all need it.

Richard Bejtlich

Director of Incident Response, General Electric Manassas Park, VA

Was this article helpful?

0 0

Post a comment