Content Delivery Networks (CDNs) utilize advanced technology that can be difficult to understand for those that are not technically savvy. For those considering adopting a CDN to effectively distribute their content, it is important to understand exactly how the infrastructure works.
CDNs are most commonly used for transferring video, audio, text and application content to users around the world. The previous method of distribution involved a centralized server that would send the information to different areas of the world. Unfortunately due to the distance, it took much longer for the page to load and the user to view the content. For instance, a user in Russia could access a U.S.-based page which would be transferred and displayed.
Optimization using CDNs
CDNs significantly cut down on page loading times and efficiency due to the technology behind the system. CDN nodes are deployed to many locations across the globe using multiple backbones. The nodes communicate with one another to fulfill all requests for content from users. The information is transparently transferred to optimize the delivery procedure.
This optimization is typically in the form of increasing global availability of specific content, reducing bandwidth costs and improving user performance through the reduction of page loading times. Optimization is an idea that was not available with the previous method due to the distance from the server to the viewer.
Optimizing by Performance
The number of nodes in a CDN varies based on the architecture of the system. In some infrastructures the number reaches in the thousands of nodes hosting tens of thousands of servers on a remote Point of Presence (PoP). In other instances there are a small number of geographical nodes across a global network.
Requests by users for content are algorithmically directly to nodes that are optimized for performance. During the optimization process, locations that are the best (highest performing) for providing the content to the user are automatically selected. Finding the highest performing location is measured by selecting those nodes with the fewest hops, fewest number of network seconds from the client or highest availability with regards to server performance.
Optimizing by Cost
On the other hand, organizations and content owners have the option of optimizing the nodes and PoP’s based on cost. In this instance, the locations that operate with the least expense are selected to deliver the content to the requestor. Therefore, performance and availability are not a factor.
In the best scenario the two options align allowing for the end user to receive the content from a node that has optimal performance and the lowest cost. These servers are typically at the edge of the network (Edge Network). The Edge Network is grown from the origin outward through the acquisition of colocation facilities, bandwidth and additional servers.
With the Internet being designed using the end-to-end principle, a core network transfer the information to network end-points and thus hosts and clients. The result of this idea is the core network becomes specializes, optimized and simplified to only send data packets of information.
However, CDNs supplement this transportation network using intelligent applications that implement methods created to optimize content delivery. The overlay utilizes web caching, server-load balancing, content services and request routing.
A web cache stores the most popular content on servers that feature the highest demand for that specific content. In other words, a popular video may be stored on a server in the Philadelphia, Pennsylvania because that server features the highest number of requests for that video in the world. Web caching significantly reduces server load and bandwidth requirements thus improving response times.
This technology utilizes global load balancing to share traffic across several servers. A switch is assigned to a single virtual IP address and traffic is diverted to a physical server attached to the switch. This improves load balancing and scalability while increasing total capacity.
Content service protocols are designed to grant access to a variety of content services across the entire network. In the late ‘90’s the Internet Content Adaptation Protocol (ICAP0 was created to develop a standard for the connection of application servers.
When simplified, the technology behind CDN is much simpler than it seems. However, since the popularity of CDNs continues to grow and the technology continues to change, it may become more complicated in the future.