charlierussell

Feeding the On-Demand Beast

Blog Post created by charlierussell on Mar 9, 2017

Reduce the Cost and Complexity of Your Content Delivery Workflows

 

We’re in an on-demand world. Video on demand, streaming audio, web surfing, online shopping, gaming and more – today’s digital audience wants the world on their schedule. They use a wide variety of laptops, tablets, phones, and TV-connected devices, sometimes in conjunction. Tens of billions of hours of content per month are consumed globally. And consumers want a great experience – instant response, high quality, no glitches or delays or interruptions.

 

Collectively, it’s a massive on-demand beast that must be fed.


If you’re a company delivering content, what does it mean to feed that on-demand beast?


The Delivery Network

You want to create a great experience for your audience. You’ve probably learned that the open Internet fails to deliver at scale. As a result, you may already use a Content Delivery Network (CDN), to accelerate delivery to your audience. One of the important advantages of a CDN is that it uses edge cache to temporarily store many copies of popular content at many sites around the edges of the network.


Feeding the Delivery Network

But what if someone requests on-demand content that isn’t stored in the edge cache? In fact, this happens millions of times per day. Edge cache generally contains popular on-demand content – hit movies and videos, new software downloads, new hit songs, popular items on web stores, images and animations from top web pages, and so on.


In practice, the majority of a typical company’s content library isn’t stored in the edge cache. It’s stored on one or more file servers or video servers – origin servers. When a user requests an asset that isn’t in the edge cache, the network needs to retrieve it from origin storage.


The DINAS of Feeding the Delivery Network

How quickly and reliably can you fetch content from origin storage? It depends. Here are some factors that can make a dramatic difference in user experience – I call them the DINAS:

 

1. Distance from the user 
The earth is a big place. If your origin storage is halfway around world from the user, one-way delay could be 300 milliseconds or more – each way. That means the user waits over a half-second just for request and retrieval from origin. You can deliver a much better response time if you have multiple origins in multiple regions, located close to where your users are.


2. Integration with the delivery network
Traffic flow is hampered if it has to traverse different networks, locations and links between the origin and the delivery network. You can streamline and accelerate traffic flow if the origin storage is tightly integrated with the delivery network, especially if it’s collocated in the same sites as edge devices.

 

3. Network from the origin to the user
The open internet has many experience-impacting obstacles. It’s subject to link congestion, node congestion, rerouting, timeouts, retransmissions, and other issues. And of course, the farther your content must travel over the open Internet, the more issues it can face. The good news is you can bypass this with a high-bandwidth private fiber network.

 

4. Availability
There’s simply no way to deliver content if it isn’t available in the first place. Outages can be caused by hard drive failures, network device issues, local power failures, regional network issues, general network congestion, and more. You can ensure availability with redundant origin servers in multiple sites. That way, even if there are issues in one origin site or one network region, the network can fetch the asset from somewhere else.

 

5. Software intelligence
If you have intelligent software that can automatically choose the fastest origin for each user session, you’ll consistently deliver the best possible response.


Feeding the Feeder

Uploading content into your origin storage, and replicating it for multiple regions, can cost significant resources, time, effort, and money. If you have to do numerous manual operations, or develop custom code, or spend time on asset tracking and management, you’ll be inefficient and burn up your money, and burn out your people.


Extra fees are another way to burn through budgets. Low cost storage-at-rest can be tempting, but extra fees can dwarf the cost of basic storage. Depending on your vendor, extra hidden costs can include fees for uploading content, copying it to multiple sites, moving it, retrieving it, transferring it to your delivery network, even requesting a directory list.

 

The Bottom Line

Whether you’re delivering video-on-demand, over-the-top (OTT) video, file distribution, gaming, web acceleration, e-commerce, or other content, your origin storage is more important than you may realize.


The optimal choice for origin storage is likely to be designed specifically for content delivery. Flexible workflow will make your job easier. Automated replication to multiple sites ensures delivery wherever your users are located. High speed design and integration with the delivery network will give your users fast response and consistently high quality. Intelligent software will serve content from the fastest site so you delight your audience on a global scale. And of course, you’ll want to deliver great user experience on budget, without hidden fees.


And that’s how you feed the on-demand beast.

Outcomes