Buyers' Guide to Content Delivery Networks 2017
With constant rumors about the death of various streaming technologies—from Flash Player (mostly false) to RTMP (false to Silverlight and Windows Media (true)—there seems to be a corollary assumption that there is no longer a need for content delivery networks (CDNs).
The sentiment is understandable, especially after the last few years of HTTP-centric streaming media delivery—with their use of dynamic HTTP delivery, coupled with adaptive bitrate (ABR ) intelligence— have led to a perception that CDNs aren’t necessary for MPEG-DASH or Apple HTTP Live Streaming (HLS).
To paraphrase Mark Twain, it turns out the death of the CDN has been greatly exaggerated.
Additional innovations in video streaming delivery via CDN continue apace, with additional standards emerging as over-the-top (OTT) streaming media continues to grow audience sizes.
There are several standards bodies that concern themselves with the inner workings of CDNs, including the International Organization for Standardization (ISO), International Telecommunication Union (ITU), and Moving Picture Experts Group (MPEG). In addition, the Internet Engineering Task Force (IETF) has issued a number of frameworks, including subsequent specifications, as part of key proposed CDN standards.
What Is a CDN?
To help readers of this Buyers’ Guide, the IETF definition of a CDN may best explain the continuing necessity of these networks. The IETF defines the CDN as the sum of its parts, operating at the mid-tier layers of the OSI networking model.
“A CDN is an infrastructure of network elements operating at layer 4 through layer 7, arranged for the efficient distribution and delivery of digital content. Such content includes, but is not limited to, webpages and images delivered via HTTP, and streaming of continuous media delivered via HTTP, RTSP, RTMP, etc.”
While there are enterprise CDNs that focus on delivering digital content for a single company, most CDNs provide services to multiple content owners or other service providers.
How Do I Price a CDN Service?
CDNs price their services in various ways, with so many different variations that it’s not practical to describe them all in this Buyers’ Guide.
But there are five prices to consider: egress bandwidth, ingress bandwidth, storage, geography and regions, and HTTP requests. We will look at these based on anticipated cost to a content owner, from highest to lowest typical cost to a CDN customer.
This is the highest cost for content owners, and it is often priced in terms of gigabytes (GB) delivered over the course of a month. In addition, for live streaming, the amount of aggregate bandwidth and number of simultaneous viewers are also factors to consider for pricing.
Pricing has traditionally been based on a tiered model, where content owners pay different per-gigabyte pricing for different tiers. This approach allows CDNs to charge a higher overall rate for the first few tiers, with those customers using less than 10TB (terabytes) per month paying the highest rate.
According to at least one CDN, though, the tiered model isn’t the best pricing approach for its customers. MaxCDN says that it rejects the “stair-step pricing model used by companies like Amazon for its CloudFront CDN, in favor of a flat-rate pricing model.
“With the flat line pricing model, you pay the 500TB per-gigabyte price for ALL bandwidth used,” a MaxCDN tutorial explains. “If you fall in the 500TB tier with MaxCDN, you pay $0.03/GB for every gigabyte used, regardless of what tier it’s under.”
A limited number of CDNs also charge for the bandwidth required to upload content to the CDN. This is known as ingress pricing.
The use of ingress pricing has gone in two different directions in recent years. The first, used by CDNs such as Highwinds, is to price the cost of ingress bandwidth at the same price as egress bandwidth, using 95th percentile pricing for both.
A second way to address ingress pricing is to completely eliminate the cost. For instance, recent pricing documentation for the Google Cloud platform indicates that “traffic between Google Cloud Platform and pre-approved CDN Interconnect locations is ... free for all regions.”
What’s a CDN Interconnect (CDNI) location? We cover that in the final part of this Buyers’ Guide.
This is often the second-highest charge for content owners. Some CDNs differentiate their pricing for storage between push and pull. Just like an origin server stream is pulled by a CDN and then replicated, a pull zone in CDN terms is one where the original web content resides on the customer’s web server and is retrieved by the CDN and cached for a specific time frame. A push zone for a CDN is one where the customer actively replicates its web server on one of the CDN’s servers. Push CDN zones can be populated by FTP, SFTP, or via REST-based APIs.
CDN-hunters, read this first for an explanation of the basic key performance indicators to explore in order to make the right choice for any size company.
It takes serious investments to stream to all viewers on all devices. For many content companies, the costs will be too high.
Researchers are creating algorithms that remove non-essential information from a video, allowing viewers to watch it faster than in real-time.