Video: What Defines a Successful Media-Based App?
Watch more Streaming Media Short Cuts videos.
Read the complete transcript of this clip:
Jamie Stackhouse: Let's go through an exercise of defining what contributes to success of your media-based application, where we really set up the performance and the cost as the balance. As a business, generally you're looking to minimize costs, so that's one of the business requirements. On the other side, you're also making sure that people continue to use the application, so perhaps, let's say we're a VOD platform, and we wanted to make sure that we were performing the lowest-latency streams, just because that was something that, say, the technological implementer behind the platform wanted to achieve. Might not be the best scenario. You can make it happen, but I bet that, as a business, you had a budget in mind.
So, to minimize that cost, providing the service, well, we do it, but we do it by essentially running very, very small, virtualized services, and it's only going to cost us a magnitude of like $50 a month. Well, we get a bunch of feedback from everybody who's using our application, that the screens are all freezing. We go investigate, and it turns out, well, all of the streaming solution is pinning our servers at 100% CPU, and they just can't keep up. We're using a stateful protocol.
So, we throw some capital, traditionally, at the problem. We co-locate some Edge servers, we start to build out a mesh network to deliver the streams, and then we also buy up some bandwidth, to create some direct connect traffic, to keep everything snappy.
Suddenly, our content is taking off. We're hitting hundreds of thousands of viewers, millions of viewers, and we're serving it all via this stateful protocol, to achieve the low latency, meaning that our hardware bill, or the cloud bill, potentially, is getting into the not-insignificant range.
So, we don't have the upfront cash right now, so we start looking at switching into a stateless protocol, and we determine that our users, well for the use case here, the video on-demand, they could actually deal with a couple of seconds of delay from the live Edge, so we move our cost into an OpEx model, paying for the bandwidth through a CDN partner, and we start to push millions of streams through it, and hopefully, the revenues cover the costs, but most importantly, we made the decision that it worked for the end consumer.
The example that I just went through, you know, it could sound like any number of VOD platforms that started and then ended, and some of them have continued, and they just really, at the core, all wanted users to come back and continue to watch their content. They wanted them to continue to be engaged with their content.
Amazon's Keith Wymbs and Jim De Lorenzo discuss how they've met the challenges of improving latency and time-to-first-byte to serve the millions of viewers who are tuning in to Amazon's Thursday night NFL broadcasts in this keynote from Streaming Media West 2017.