Air-traffic controllers coordinate access to runways. Busy airports have more planes wanting to take off and land than runways available. Runways can only facilitate one plane at a time. In generic terms, the air-traffic controllers coordinate access to a limited resource (the runway).
Such is the purpose of content switches: coordinating access to limited resources, in this case, through load balancing. The primary example for content switching is the load balancing of HTML and XML traffic among multiple web servers.
The number of users needing to access a particular website might exceed the capacity of a single server. You could implement multiple servers to handle the increased load. However, because every device on a network requires a unique IP address, how does the user select which server to access? Some server farms (such as large Internet-based retailers) might require hundreds of servers to accommodate their online customers.
Content switches resolve this problem by transparently intercepting all traffic destined to a website and distributing each request to an available server. For example, when a user visits http://www.cisco.com, a load-balancing content switch can ensure that a server that is not overloaded handles the request. The user sees a single web server, www.cisco.com. But in reality, there might be tens or hundreds of servers.
Content switches can provide relief in the following situations:
Server load balancing—. As described earlier, a content switch can ensure proper load balancing across a group of servers transparently to the user.
Firewall load balancing—. As server farms increase in size, companies want to protect the systems from unauthorized access. They do so with firewalls. Content switches can enable load balancing across multiple firewalls.
Distributed data centers—. As the ultimate show of redundancy, some companies might choose to duplicate all network and computing services in a data center on an alternate site. Thus, if the original data center becomes unavailable, all services can continue on the alternate site. Although the implementation of distributed data centers requires careful planning and integration, content switches become the traffic cop during an accident. Any traffic destined to a data center transparently meets the content switch first. The content switch then redirects the traffic to the original data center, as usual, or to the alternate site if the original site disappears.
At-A-Glance—Content Networking
The near instantaneous availability of data from internal and external network sources has helped transform networking from a high-tech tool with narrow appeal to a broad method of delivering business tools, educational resources, and entertainment to a worldwide audience.
Although the massive decrease in the cost of bandwidth (dollars/kilobits/seconds) has been a big part of the increased use and availability of the Internet, the network still has far too many bottlenecks that you must overcome to make networks and data distribution work efficiently.
Content networking solves this problem by mitigating the effects of network bottlenecks by providing an efficient (and transparent) means for distributing all forms of downloadable media (content) across networks and balancing traffic loads over multiple devices.
</division><division> <title>What Are the Problems to Solve?</title>The two keys to content networking are overcoming bottlenecks and maintaining relevant data across multiple sites.
A bottleneck is the slowest point in the network, which forces other faster parts of the network to perform at slower speeds than they are capable of. Although you can easily address bottlenecks by adding more bandwidth, it is not always a viable solution. In such cases, you can still address them by getting data as close to the user as possible prior to a user request. This process is known as caching.
Assuming that multiple sites are prepopulated, the issue becomes one of updating and synchronizing remote sites because many types of data have a “shelf life.”
</division><division> <title>The Brains of the Operation</title>Two key pieces of equipment are responsible for the intelligence behind content networking:
Content manager—. Responsible for the distribution (caching) and update of material. Whenever possible, it delivers content prior to a request and then, based on administrative input, performs caching updates prior to shelf-life expiration.
Content switches—. The traffic cops of the network. Whether you are accessing a website, data server, or remote network, a content switch directs you to the device that provides the shortest delay possible.
In many cases, an original source of data or information is not local to all users and must traverse a wide-area network (WAN) to reach a user. The point where a LAN ends and WAN begins is the network edge. The network edge is almost always a choke point, meaning that the available bandwidth on the LAN side of the edge is larger than that on the WAN side. To avoid annoying delays in responsiveness for the content requested, as well as for other competing traffic such as voice or video, a content manager can prepopulate a remote location with content, update a local server with real-time updates, and perform complete backups during nonpeak hours.
Remote servers populated with content from another source are often called mirror sites.
</division>At-A-Glance—Content Availability
The content manager requires some intelligence so that it can make decisions about where to prepopulate, what to prepopulate, and how to bill and provision services.
</division><division> <title>Content Routing</title>When a user requests content, a router receives the request and determines the best possible site to route the request based on the location of the user, the availability of content among different sites, and network status. Other rules defined by a network administrator might apply as well.
</division><division> <title>Content Availability</title>One of the main advantages of content switching is the ability to transparently deal with network or hardware and software errors. If a server (or a device such as a firewall) goes down, a content switch simply removes it from the list of available endpoints, and user requests are directed elsewhere. This process happens without the user realizing it. For retail websites, this transparency is especially important because users are less likely to shop on a site if availability is questionable.
</division><division> <title>Content Switching</title>The purpose of content switching is to provide intelligent load balancing for servers and other devices such as firewalls.
The need for load balancing stems from the fact that a single device cannot always adequately serve all requests at any given moment in time. Therefore, the information must reside among several devices.
Content switching prevents users from knowing the name or location of every device. Users simply request the access (typing a single URL, for example), and a content switch chooses the best device to access based on the type of request, the priority of the request, and other defined rules.
</division>