The State of Media Servers 2015
Media Servers Have One Job
Vietor says the final benefit to a media server is all about what it is not, rather than what it is.
“Media servers will have to focus on the areas where they are better than web servers,” Vietor says, “and not try to beat web servers at their own game.”
“For bare-bones adaptive streaming of a single title, using a single ABR technology like HLS,” Knowlton says, “you could put thousands of preencoded, pretransrated,
prechunked, prepackaged HTTP files in a directory with a manifest and deliver that title from a web server. But it’s not an elegant, flexible, nor feature-rich solution.”
Innovating Beyond Competition to Scale
The issue faced by any form of HTTP or media server in today’s environment is one of scale. All the current media server companies see themselves as innovators in an era where video is being adopted at a massive rate.
Cisco anticipates IP video traffic will triple by 2018, led in large part by video IP delivery on the internet, which it anticipates will grow 400% over the next 3 years. And many of those actively engaging in testing the waters for IP video delivery are what we might term “new streamers,” or those who’ve never before been tasked by their companies to delivery video over IP to other employees, customers, or even the public at large.
“Many of these new streamers have short deployment deadlines to achieve,” Knowlton says. “For some of them, this means starting with a free or low-cost video delivery
service is an acceptable first step. Unfortunately, you typically lose some level of control over key elements, such as advertising, custom branding, access control, and content protection.”
This trial-and-error approach with low-cost or free solutions is often a blessing in disguise for media server companies. Once these “new streamers” get a few months of hands-on familiarity with what low-end services offer, or don’t offer, they often begin actively seeking a solution that offers more control over both content and user experience.
In addition to offering the basics, media server companies are constantly innovating. DDVTech’s Vietor says that one goal with MistServer is to make any content playable in even older media players, including those perhaps not necessarily intended to stream content for formats that were not considered to be streaming container formats.
DDVTech has been working on an innovative way to stream live content in an MP4 container format, allowing legacy media players that have the ability to access a URL to stream live content. The URL access was intended for MP4 files, which contain end-of-file information, but MistServer forces the legacy player to accept the fact that the file is of an unknown file size, meaning that the live stream could last an additional 10 seconds, 10 minutes, or even 10 hours.
“The byte-range/millisecond position calculation is how MP4, TS, OGG and FLV streaming work in MistServer,” says Vietor, adding that other protocols such as DASH and HLS work similarly, but instead of a single virtual file per client, they use many, many virtual files per client.
In addition, Vietor says that new protocols, such as WebRTC, offer a chance to innovate even further. One reason for this sentiment, at least when it comes to streaming, is that WebRTC requires a media server for particular types of delivery.
“In order to modify WebRTC capabilities to display live/VoD video streams, we generate secure RTP packets over UDP,” Vietor says. “That’s hardly a web server task, since HTTP servers are designed to deliver packets over TCP and not UDP.”
Trending Toward Live, Mobile, and Cloud
As discussed in the “State of Mobile” article, mobile video consumption is on the rise. And it’s not just on-demand mobile content that is trending upward; recent reports show that live streaming of news and sports is leading the way on mobile devices.
In addition to mobile delivery, while laptops and desktops clearly still retain the overall edge in terms of premium online media consumption, gaming consoles and tablets are also rising in popularity for premium content delivery.
What this means, in terms of media servers, is that live streaming to many more screen types will continue to be a challenge, especially when it comes to live content delivery.
The trend for at all these new screen types, with the possible exception of enterprise content—as noted in our “State of the Enterprise” article—is toward cloud-based delivery.
“Cloud video infrastructures are going to become the default deployment option for many companies,” says Knowlton. Microsoft’s Azure, Amazon Web Services (AWS), Google Cloud Platform, and Rackspace are all participants in the Works With Wowza program.
Cloud infrastructures allow those who want to try out a media server to do so, at minimal commitment. If the intent is to stream content to a public, internet-based audience, cloud infrastructures have the added benefit of allowing a “new streamer” to test out a deployment at a much lower infrastructure impact. In many cases, the only hardware capital expenditure would be an encoder; it doesn’t even have to package into an ABR format such as DASH or HLS, since many media servers can receive a single stream in RTSP or RTMP form, then do the packaging for ABR delivery as part of the media server solution.
It’s clear that the technology problem-solving that occurred over the last decade was merely a foundation for bigger and better things in the broadband media delivery industry. Whether it’s multichannel delivery of entertainment content or just an enterprise moving toward better delivery of its CEO’s all-hands meeting, the goal is to make the technology work for the solution, rather than hinder the delivery solution.
“Microsoft never monetized media servers directly, making Windows Media Services and IIS Media Services available as free downloads for Windows Server customers, and it is now turning much of its focus to revenue-generating cloud-based services,” Knowlton says. “With the proliferation of devices and HTTP streaming formats, Adobe is moving from general-purpose media technologies to a targeted solution for broadcasters and MVPDs such as NBC and Comcast.”
In other words, now that we’ve solved the basic media delivery problems, these companies have been able to shift focus toward their core competencies, using these technologies to expand existing business lines. Yet, just because they’ve shifted back to core business focus, this doesn’t mean that media servers are dead. And that’s good news all around, as IP video delivery continues to grow and will rely on additional innovations to get to TV-sized audience scale.
[This article appears in the 2015 Streaming Media Industry Sourcebook.]
Companies and Suppliers Mentioned