Sometimes, the problems we experience with computers are a result of a legacy design. Hardware or software might have been architected 10 or 20 years ago when the world of computing was vastly different from the way it is today. As a result, the product in use today isn't as effective as it could be because of aging or obsolete design.
Such is the case of the traditional firewall, whose design dates back two decades to the late 1980s. Early firewalls consisted of packet filtering software that inspected all traffic coming into and going out of a network. If a packet of data met specific rules, its transmission was simply dropped. Later generations of firewalls were engineered to approve specific applications or to look for Internet traffic using specific ports. These legacy firewalls were built on the assumption that an application would respect its protocol which would respect the port. For example, Port 80 must mean HTTP and that must mean Web browsing. Or, Port 25 must mean SMTP and that must mean e-mail.
That's not so true today. Many modern applications are built to be flexible, meaning they change ports as needed to deliver their content. Skype and BitTorrent, for example, hop around and use multiple ports like Port 80 or 443. A traditional firewall isn't expecting this kind of traffic there. Unfortunately, the old assumptions about port mapping applications are out the window today.
This has created a cottage industry for other "bolt-on" security applications like intrusion detection/prevention systems and antivirus/antimalware scanning. These applications are meant to catch the problems that legacy firewalls sometimes miss. The result can be a patchwork of security applications that scan traffic multiple times and add to the complexity of your infrastructure.
Seeing an opportunity to "fix the firewall," a group of security engineers started a company in 2005 to redesign the firewall architecture from the ground up. These engineers took their expertise from working at places such as Check Point Software Technologies, Juniper Networks and NetScreen Technologies and started Palo Alto Networks.
They set out to design a single firewall appliance to address three business problems:
1. Identify and control applications, including enabling applications that can be productive to the business.
2. Prevent threats from harming the network.
3. Simplify the security infrastructure.
The Palo Alto firewall uses a unique single-pass process for traffic classification, user/group mapping and content scanning. Three technologies embedded into one appliance eliminate the need for bolt-on products:
* App-ID is traffic classification technology that determines the exact identity of nearly 900 applications flowing across the network, irrespective of port, protocol, SSL encryption or evasive tactics.
* User-ID links IP addresses to specific user identities, enabling visibility and control of network activity on a per-user basis. It integrates with your Active Directory to harvest relevant user information such as role and group assignments.
* Content-ID provides content scanning that prevents threats within permitted traffic, and provides granular control of Web surfing activities as well as file and data filtering.
As a result, the firewall allows you to have a fine grain policy that covers a user or group, an application and specific content all at once. For example, you could enable just the sales department to use WebEx but not the desktop sharing feature of the application. Or perhaps you want to allow the marketing department to access social media applications like Facebook, but no one else.
You would think that asking a single device to do all this inspection and classification at once would create latency. But remember, the engineers "fixed" the firewall by rethinking the architecture. They built this firewall with function-specific hardware that enables parallel processing. Instead of packets doing multiple passes through various functions, the data streams are processed in an essentially linear model. This helps to optimize performance, even in the face of massive volumes of traffic.
As for using the Palo Alto firewall to simplify your infrastructure, it's possible to replace multiple devices with just this one. That's what John Kovacevich, systems analyst at Texas A&M University at Galveston has done. After deploying his Palo Alto Networks box, he retired his old traffic shaper and now uses his new firewall to manage traffic bandwidth. "We can't stop students from using peer-to-peer applications, but we can shape it so that it gets a lower priority and less bandwidth than our e-learning applications," according to Kovacevich.
The university also uses the Palo Alto device to protect its network against viruses and other threats. Because of the integration with Active Directory, Kovacevich is able to identify the specific student whose machine seems to be infected. He can call the student and schedule a cleanup of the PC. "I don't have to track down someone by the port anymore. Now I can do it by name," Kovacevich says.
He likes that he now has just one device to manage for multiple functions. "I had multiple products before and now they are just sitting on a shelf," Kovacevich says. "Palo Alto offers a product with a lot of features in one box at a very competitive price."
Those words are music to the ears of the product engineers who took a chance on designing a new kind of firewall using best practices for the modern era of Internet applications.
This story, "Engineers fix the shortcomings of the traditional firewall" was originally published by Network World.