ECDN 101: Building Blocks for Enterprise Content Delivery Networks

ECDNs: Caching and Beyond

A lot of people assume that you can position a number of caches out around the edge of the network, and presto: instant ECDN. But building an enterprise CDN requires more than selecting caching technologies. It requires a solution encompassing aspects of both distributing and accessing rich media.

Still, Web caches are a key component. A Web cache is used to store--and serve--Web objects (objects that would normally be served from a Web server). A Web cache monitors Internet traffic, intercepts requests for Web objects and then fulfills those requests from the set of objects it stores--instead of those objects being served from across the Internet by a Web server. If a cache receives a request for an object that it does not already store, the cache forwards the request to the origin server. The origin server sends a copy of the object to the cache, and the cache saves a copy of the object in its object store and sends the object on to the requester.

Caching also reduces capital and operating cost. The primary reason Web caches are so popular is that they save time and money. Another major reason for caching's popularity is that Web caches are flexible and can be deployed at the data center core or the network edge. In either location, Web caches reduce network traffic, conserve server resources, and accelerate content delivery.

Caching conserves bandwidth. Deployed between a client's browser and the Internet, caches can significantly reduce wide area bandwidth consumption by locally storing objects that would normally be delivered from a Web server at corporate headquarters or even somewhere across the Internet. This close-to-the-client deployment mode is commonly known as forward proxy caching. Forward proxy implementations can reduce wide area network traffic by 30 to 50 percent (results vary based on the "cacheability" of the requested content), allowing more bandwidth for higher priority traffic.

Caching also acts as a point of control and security. Today's caches frequently include support for content filtering, anti-virus, access control and bandwidth management. Anti-virus and content filtering give users an extra level of security across the network. Caches provide a logical point in the network for this type of functionality to take place and can act as either a primary or secondary point for this type of security. With the increasing load on the enterprise networks, the access control and bandwidth management further assists in the reduction of overall network utilization by making sure that only approved users get access to bandwidth, and that the bandwidth is being allocated according to the business objectives.

Delivering Content To the Caches
The second key element of an ECDN revolves around its ability to manage the content. That is, you have to figure out ways of taking that content (which is in your data center) and delivering it out to all the different cache appliances so it's available to the end users, quickly. You have to make sure that the content is synchronized so that everyone is getting the same information at the same time--essentially making sure the content is fresh and up-to-date.

Up to 30 percent of cache-initiated traffic on your network are checks to see if requested items are still fresh. You can eliminate the need for freshness checking by ensuring that your content is always fresh. You can achieve this by pre-populating, or pushing, content to the cache before your audience request it. When you automatically push a new, or updated, Web object to a cache, you are guaranteed that the content in the cache is always fresh and there is no reason for the cache to initiate a freshness check. The job of the cache is streamlined to serving requests, the wide area network is used only when necessary, the impact on the Web server is reduced, and your audience gets an improved browsing experience.

Pre-populating content takes on even more importance with broadband content. The sizes of rich content files can be huge and they are increasing every day. Compression technologies have been invented specifically for these new and emerging types of media. Streaming audio and video files are created with specific line speeds in mind. Lower speed lines means lower quality audio and video. If a user with a 56k Internet connection requests a video file, they will receive a version of that file with poorer image and sound quality than a higher speed connection would deliver.

However, with content pre-populating you can push a high-resolution file across low speed lines directly to your Web cache in the branch office and then serve that streaming file at the top speed of your LAN. For example, a branch office with a 56k wide-area connection can enjoy the higher quality of a video file created for a much faster connection. Audiences experience the highest quality video while your company's bottom line enjoys the reduced cost of a lower bandwidth connection.

Load Balancing/Traffic Management
Next, ECDNs require traffic management and server load balancing products. Those Internet traffic managers keep traffic moving. Their job is to ensure that each and every packet reaches the destination that will do the best job of fulfilling that packet's mission. For content delivery and acceleration, best is generally defined as the cache or server with the fastest response time (and with the correct content, of course).

Wide-area load balancing is the traffic management "secret sauce" that many of the current content delivery network (CDN) services use to direct audiences to the "right" cache. Wide-area load balancing can be effectively used by enterprises in your private ECDN implementations as well, especially implementations with multiple data centers or cache hierarchies.

In either case, multiple data centers or cache hierarchies, the wide-area load balancer directs a request from a cache at the network edge to either a data center or a parent cache. The primary value of wide area load balancing is ensuring that an available and high-performance device will fulfill the edge cache request.

Caching, content redirection technologies (local and global redirection/load balancers) and content distribution and synchronization functionality then are the basic building blocks for an ECDN. In a future class I'll talk about the other additional components that can help further increase ECDN performance, scalability, reliability and security.