What you need to know about On Demand Video Delivery
Hello everyone. This is Charlie Kraus from Limelight, and I'm going to be co-presenting with Ernest Russell. We're both from the product marketing side, focused on Limelight's video solutions. So this webinar is the third in our bit by bit series. And this will focus on a video-on-demand workflows and how to deliver the best possible quality video on demand to your audiences. So let me just go through the agenda of this particular webinar first. So we're going to cover a little bit of market data, as usual, just to get you up to date on what things are like out there. And then talk about supporting the various devices that the audience is used to view some video on demand, a workflow options that you have to deliver. Managing video is very important, as well as the monitoring and analytics, which goes to monetization and just understanding behavior and then a brief summary at the end.
So from our recent online video surveys that we take every year, here is some of the data that we're seeing. So every year we see the total number of online video hours that people watch going up for the first time and taking the survey over like five years, we've seen that smartphones and tablets, and rather, you know, generally mobile devices have taken over is the most popular way to view over a PCs and computers. And that was that just recently changed. Always quality rebuffering is, is a bad thing. And audience patience is, you know, very slim. So we see here by the time you get to like to rebuffering, you've basically lost a huge chunk. Almost 70% of your audiences have clicked off in terms of what people watch. The various types that video-on-demand content certainly is significant, almost dominant share, new sports and other live content is also growing in popularity, but the big chunk of online video is still video on demand. So with that, I'm going to turn it over to Ernest, and he will take you through video-on-demand workflows.
All right. Thank you, Charlie. So as Charlie was saying, multi-device delivery, delivering to the different devices that are out there has to be a consideration when you're building your VOD workflow. You want to make sure that you are prepared to deliver to OTT devices to smart TVs, tablets, phones, laptops. I mean, just in the range of phones, there's so many different devices and different types out there. So you want to be prepared for that by transcoding or transmuxing, converting your video to flavors that are made for streaming and video delivery. And so let me take you through step by step on how we, how your video gets from your camera to the end viewer.
So here's the video-on-demand workflow. Step-By-Step first, starting with video ingest, where you capture the video and upload it. So this is the step in which you're recording a video on your camera. And so usually the next step is the conversion of that video into the various flavors for the different devices that we want to deliver to, and also preparation for adaptive bit rate and that's to scale, not just the size of the screen, but also the bit rate for various network conditions. And so after the video conversion, then we go into what we have as step three here, video storage and delivery. And so that includes storing and distributing video across the network. And of course, video playback could be various devices and smartphones. And then, finally, you want to make sure that you have video analytics to find out where and how your viewers are watching.
So let's take a look at the complete workflow here. And so here's starting on the left. You see that we have ingest where again, we're taking the video that was captured live and ingesting that on-demand-video for later playback. And so, after we ingest that video or upload that video, it goes into conversion, and that's where we create the different flavors that are needed. Some of the most popular are HLS and DASH. But we also see a lot of impact delivery for VOD. And so, storing and delivering devices to various devices requires strategy. We want to make sure that we deliver from as close to the viewer as we can, and that way, the viewer can get the highest quality playback experience. And so that brings us again to the analysis part.
And so we want to make sure that in our video workflow we have the feedback of our users and how they're watching, where they are watching from. One important thing, when we're considering the VOD workflow is also a video and metadata management. So things like titles, audio tagging, timing, those are very important when we're building a VOD library because we want to be able to go back and search for the videos and also be able to do things like preview and maybe arrange them in a playlist or a channel. And so we want to make sure that you consider video and metadata management upfront and tools that will help you do that. Sometimes that's part of the ingest process, and sometimes it comes later in packaging.
So let's talk in more detail about step one. And that was again, getting video into our VOD delivery system, and some popular ways of doing that are FTP. Sometimes, you may drag and drop through a GUI, API upload or even accelerated upload through Asperra. And so some best practices when you're ingesting or moving your files into storage is you want to ingest into two places as good practice and that way you have a backup in case one goes down, that can be one primary, one secondary, or both, both active, but you want to have that redundancy in place and also file replication that scatters the files close to where your viewers are watching.
Moving on to the next step, here is the conversion of the video files. And again, that entails the preparation for adaptive bit rate the IDR alignment to make sure that transitioning between different bit rates and network, when your networks speed changes is a smooth transition and here is where that packaging and the metadata comes in, that we talked about things like adding a subtitles audio shuffling, modifying manifest, and also you want to make sure that you have your quality control here and make sure that the assets that you are in coding are what you want you to be wish to see.
Moving on to the next step here is storage and delivery. And so you want to make sure that you're prepared to deliver to a large audience, right? And so some of the challenges to doing that on-premise are, for one, it takes the staff to monitor and support, and it's tough to scale, especially in the situation where you're going viral, right? So the CDN provides value there. So where it protects your origin server from being overloaded when a video goes viral, also there's the performance that you get from the video being cashed close to the user.
And so when you play by client receives your video through, it'll also receive a manifest and requests chunks through that. And here's where we'll get into more detail in our security webinar coming up. So stay tuned for that one. We'll talk more about how the client, whether that's a smart smartphone or tablet, receives the manifest and also where DRM applies here.
And finally, we want to make sure that we have monitoring and analytics, right? That we know where our viewers are watching and what the viewer devices they're using. Even things like browser information can become important to help us make informed business decisions, right? How long are viewers watching? What's the average time viewed or the percentage watched. And so that I'll hand it over to Charlie now to wrap up.
Thank you, Ernest. I hope you had the impression watching this that there's a lot of flexibility in these video workflows that can pretty much solve all the challenges you have getting that content out. We know that viewers increasingly watch from a range of different devices, and this changes over time. So rather than have to worry about this, CDNs have the ability to detect exactly what kind of devices out there determine what format video they want to receive and take care of that on the fly, transmuxing to get it to them. And as devices change over time and they get updated CDNs, keep up with that information. So it relieves you of a lot of the burden of that. You have a lot of options in terms of how much you want to prepare the video and how much you want to leave to CDNs, but, you know, by offloading, most of that, it takes a lot of headaches out of that.
Again, we have these great tools to do the transcoding, transmuxing and get the right bit rates out and just generally deliver the best possible quality to the audience. And monetization is always an issue. So having those analytics about viewer behavior, how long they're engaged, what they're watching, how many times they come back, how long they stay on a particular video is all critical information. When you start negotiating ad rates and so forth and, and even understanding what kind of content they like to watch in terms of building your library in the future.
So storage was mentioned a few times, and what I want to do now is put in a plug for the next video. So following this up in a, in a good, a series is optimizing video storage. So we put it in this order. So in number four, the various mentions of caching and moving content close to users, we'll talk about in the next webinar about how you can leverage the CDN storage to both have it replicated in multiple places for redundancy, and also ways you can automatically load that when you have very large libraries, and you worry about how would you ever get that kind of content into the storage?
There's some really cool tools that let you use sort of the viewers' market data to determine what gets stored and where it's positioned, and so forth.
So we want to thank you for joining us today. We hope this was useful to you in looking at your video-on-demand workflows, and we hope to see you on another webinar in the future. Thank you!