Today, there are sophisticated tools that let you take advantage of Layer 7 information returned by applications to end users, especially in the high-availability, loadbalancing realm. These tools let you easily verify that the site content is responsive and correct, or test your site from a customer's point of view to ensure the correct applications and content are present.
But while it's possible to verify that the right content is being sent, it's also becoming possible to open the packets of data coming across the wire (regardless of IP address or port) and make load-balancing decisions based on that information.
Essentially, this migration of intelligence moves beyond Layer 4 functionality.
For example, take Port 80. There are many types of traffic that flow over this port in addition to the normal types of Web traffic. The problem? Devices whose capabilities end at Layer 4 are blind to the various types of traffic flowing through the pipe, so they treat all traffic the same.
But all traffic isn't the same. For a load-balancing product, it would be quite useful to know whether the data flowing through the port is streaming media or simply a request for an item out of your catalog. Perhaps you would like to place a higher priority on the person who wants the catalog item. Devices that end at Layer 4 treat both types of data the same, opening the possibility of sending the streaming media traffic to a server that is incapable of providing a response, resulting in error messages and delays.
Yet the new Layer 7 intelligence gives you additional control - total traffic and content control. By giving you the freedom to fully open the application/presentation layer of traffic and closely examine what's inside, you can make more intelligent load-balancing decisions based on types of applications - not just on IP and port number.
This allows you to make complete load-balancing decisions based not only on URLs, but also on actual application types regardless of the port number they happen to be using. That would allow you to recognize, for example, a video conferencing stream, and make appropriate load-balancing decisions based on that information, even though the application may be using a dynamically assigned port.
Part of the functionality of this type of Layer-7-aware product is to guarantee that different types of traffic can be assigned different priority levels.
Instead of relying on routing equipment or applications to identify traffic through Differentiated Services (Diff-Serv), Common Open Policy Service or other quality-of-service protocols, the Layer-7-aware device can sift through the traffic and assign priorities itself. This frees you from having to rely on the application or the network gear for these purposes.
Are there any set standards for this type of Layer 7 functionality? The short answer is no. Layer-7-aware functionality is more complementary - it lives in harmony with networks that offer things such as Diff-Serv. It takes generic traffic and says, "This traffic requires this type of service bit to be set because it's voice-over IP, yet this other type of traffic needs a low-priority type of service bit."
The big news, however, is the final benefit that these types of devices present. In the past, there was always a trade-off between intelligence and speed. With Layer-7-aware technology, you'll be able to make more intelligent traffic decisions at wire speed. You'll be free to make informed decisions on various types of traffic and its destination, optimizing your Web access and providing a better end-user Web experience.
The bottom line is efficient traffic prioritization and intelligent load balancing.
Matte is director of product management at F5 Networks, a maker of Internet traffic-management and content-management products in Seattle. He can be reached at firstname.lastname@example.org.