Video: Peer-Accelerated Delivery for Large-Scale OTT Optimization

Tim: Welcome back to Streaming Media East 2017. I'm Tim Siglin, contributing editor on Streaming Media magazine and media strategy principal at Reel Solver, Inc. Today, we have Erica with Streamroot. Erica, introduce yourself to the audience.

Erica: Hi. My name is Erica Beavers. I am heading marketing and partnership initiatives at Streamroot from our French office in Paris.

Tim: Tell me what exactly is Streamroot, because several people who I've known and trust have come up to me and said, "Streamroot's got something interesting going." What do you do?

Erica: Streamroot is a provider of OTT optimization services for large-scale OTT broadcasters both live and VOD. What we've developed over the past several years is a peer-accelerated delivery technology for large-scale OTT broadcasting. By peer-accelerated, I mean that we have our core solution which is called Streamroot DNA for Distributed Network Architecture, creates a sort of mesh network so that we're not only sourcing video segments from the CDN but also from a network of viewers that are actually devices that are watching that content at the same time.

Tim: So VOD and live?

Erica: Exactly, exactly. Live streaming is a particularly interesting use case for us as everyone is tuning in at the same time and we see there're bigger and bigger traffic spikes that CDNs tend to have a harder and harder time to actually have that load, and it's very expensive.

Tim: Given that fact and that's where say peer-to-peer would shine, what are you asking the CDNs to do differently?

Erica: Essentially what we would ask the broadcaster ... we have a couple of small configurations but we integrate seamlessly into the broadcaster workflow. We're integrated into the video player at the video player level, which is making the decision as to where to get the actual segments. But in terms of their CDN, we're totally CDN-agnostic, codec-agnostic, we work with their workflow. I'm not trying to change that.

Tim: You say you're integrated at the player level. Does that mean that it requires an app or can the player be a browser-based player, as well?

Erica: We're integrated into many of the open source and proprietary HTML5 players so we're HLS.JS, DASH.JS, we have a Brightcove integration, JW player, Bitmovin, a lot of the big names out there, we already have a kind of plug-and-play integration.

Tim: In the last five or six years, you had a number of companies trying to do peer-to-peer. You had Octoshape, which Akamai acquired. You even had RTMFP which was what Adobe was trying to do as part of the flow protocol. Issues they faced were, say with enterprise, you had to pierce the firewall with a unicast and then essentially once you're inside, "infect" ... for the lack of a better word ... on a peer-to-peer scenario. How do you all handle those kinds of things?

Erica: To address the first part of your question, how we're different from some of these legacy peer-to-peer systems such as Octoshape: Streamroot is doing this entirely plugin-free. Today with HTML5 we don't necessarily want to have a plugin anymore, and to have a seamless user experience across platforms, so we're doing this via the WebRTC protocol which is entirely plugin-free. For users that are actually behind the corporate firewall, what Streamroot has done thus far is we're very much optimized for the big public-use cases. That's where we see a lot of the demand and a lot of those really big cost and quality challenges.

Tim: Like big sporting events and that sort of thing.

Erica: Exactly. However, we do also address the corporate streaming market, and in that case what we would have to generally do is just bring our matching intelligence within the actual company and the local area networks, so that's not an issue either. The way of handling the two systems is quite different given the different network conditions.

Tim: From your unique selling position, what are the things that you take back to a broadcaster in terms of cost of deployment, return on investment, and ultimately bandwidth savings say, in lower latency? How do you approach each of those different areas?

Erica: So, in terms of the bandwidth savings, we're looking for a multi-bitrate use case with viewers all over the world. We're achieving anything from about 55 to 75% offload from the CDN. That, in terms of savings, translates to a large amount of savings of course. What we've also tried to do, and that's something strategic for us and also for broadcasters, we've come up with a fixed pricing model. We think that obviously the CDN market has become very commoditized, and that per-gigabyte product pricing model is not necessarily conducive to price stability and predictability for broadcasters, so what we generally do is offer a flat fee that can be based on a maximum number of concurrent viewers or an average number of concurrent viewers. They can stream in 4K, they can stream on mobile, whatever amount of bandwidth they need to actually stream that content, that's fine.

Tim: The price is strictly by users as opposed to by bandwidth say, or resolution?

Erica: We're pretty flexible with that, so if a broadcaster absolutely wants something that is bandwidth-based, we can do that. But we generally find that the cost predictability offered by a flat fee is a lot better.

Tim: In a typical scenario with a 55 to 75% offload off of CDN versus your flat fee, what's the return on investment?

Erica: The return on investment is really immediate and it actually outpaces the savings on the bandwidth.

Tim: So, essentially, you're able to go to the broadcaster and say, "With the offload at CDN cost, even adding in our cost, you're still going to save money on that.”

Erica: Exactly, exactly.

Tim: How long has the company been around?

Erica: We were founded at the end of 2013 in France, and are now headquartered in New York as of 2015.

Tim: Do you have VC from the US, or is it a privately funded company?

Erica: Yes, we have venture capital and a group of business angels as well from the US.

Tim: Very good. Was that one of the reasons you moved to New York, or was it because everybody likes New York?

Erica: The biggest media market in the world is in the United States.

Tim: Sure. Absolutely, yes.

Erica: We got a great initial traction in France and really said that we need to be at the heart of the industry in New York, in the United States.

Tim: Very good, and then will you expect to expand to say offices in London, Tokyo, some of the other large markets as well?

Erica: Yes, our market ... we're already present in South America and we plan to open offices in several other locations as well. We have a global market, so ...

Tim: Awesome. Erica, thank you very much for your time.

Erica: Thank you so much, and thank you, Tim.

•