|Create Date||May 20, 2015|
|Last Updated||May 21, 2015|
By Alon Maor, CEO, Qwilt
A look at the challenges facing online video delivery from the network point-of-view
The 50 year old, highly asymmetrical, model of broadcast television is being shaken at its core by consumers, technology and new service provider business models. Rivalling some of the greatest technology disruptions in history, the online video phenomenon has catapulted the market into an unprecedented – and very exciting – transformation. Undoubtedly, it was the advent of long form HD video from sources like Netflix, Amazon and Hulu that ushered in a new generation of viewers and transformed consumer viewing behaviour. The novel, early online-video days of watching clever three minute YouTube clips have been eclipsed and the notion that you once had to wait for a TV show to come on air in order of broadcast already seems laughable.
High quality video streaming is not only making its way into our living rooms, via the popular uptake of smart TVs, but increasingly onto our mobile devices too. Today’s viewers now not only expect to be able to watch their favourite TV show or film, anytime, anywhere, but also on any device of their choosing. Unfortunately, existing IP networks, already straining under the pressure of serving this volume of streamed content, are being pushed to their limits. This consumption trend is confirmed in research that estimates half of a mobile viewers’ time is already spent watching videos that are longer than 30 minutes. Interestingly, many cable operators are beginning to recognise – and, in some cases, admit publicly – that as time goes on, their broadband internet service offering, not their cable TV service, will be their strategic product line.
Faced with the inevitability of online video streaming going mainstream, it is critical that network operators move now to prepare their networks for the future of online video, or else both viewer satisfaction and, crucially, subscriber dollars will be at stake. It’s important to remember here that at every step in the video streaming ecosystem, from the content provider’s origin server, to commercial CDN, to transit and peering exchange points, to the network operator and finally to the home, a video stream must navigate a diverse group of commercial and business interests and gateways.
The challenges, therefore, are manifold; for one, any surge in traffic load during peak video consumption hours – like around the time of the season finale of House of Cards, for example – can clog up the network and trigger an annoying ‘buffering, please standby’ notification. Aside from the viewer irritation of having to wait for content to be delivered, traffic pressure inevitably impacts video picture quality and Quality of Service (QoS) for other services – including other latency-sensitive applications. Unfortunately, in a highly competitive industry, slower downloads, delays in response, lower bit rates and subpar viewing experiences simply result in consumer churn – and, on occasion, a public flogging via social media.
In the past, conventional wisdom would guide an operator to address the problem with brute force – buying more routers, switches and links to increase capacity. However, just throwing money at the problem is proving to be an approach as unsustainable as it is costly. A long-term solution for online video is not just about building bigger networks, or becoming trapped in closed, proprietary systems, it’s about building an intelligent, open network. There is broad-based agreement with the architectural principle that caching video content close to the consumer is an essential part of the overall strategy to deal with the online video problem. However, what’s missing from this conversation is the choice in caching architecture that must be made along the way. This choice and the implications that flow from it are critically important and worth a deeper discussion.
The notion that there is a choice in caching architecture is often obscured today by the ongoing deployment of content-specific caches. These closed and proprietary caches handle popular video traffic but solve the problem for only one content provider at a time. For example, the Google Fiber team describe their support for various closed cache systems in a recent blog but do not address the option to deploy an open cache architecture which would benefit all content providers and consumers. Ultimately, when addressing a future in which the content provider landscape is both diverse and dynamic, ‘open’ simply beats ‘closed’. Moreover, the imperative to build an open architecture for streaming video appears to be an essential part of the recent Net Neutrality ruling from the US Federal Communications Commission. This ruling affirms the ‘Open and Free Internet’ and clearly requires the US Internet Service Providers, for both mobile and fixed broadband networks, to treat all content without preference. As this ruling is fully implemented, any content-provider-specific cache deployed inside the last mile network will likely be called into question as preferential treatment that is prohibited by the new FCC rules. Further, we expect the EU and many other regulatory bodies around the world to adopt much of the FCC’s net neutrality doctrine.
Support for the open caching movement is growing fast as evidenced by the announcement last fall of Streaming Video Alliance whose 17 founding members have made clear the need for an open architecture to allow online video to flourish. To be sure, open caching of content will eventually become part of the core network infrastructure just as routers and switches are core today. Indeed, we already have the open internet, TCP/IP, routing and switching as shining examples of how open technology as infrastructure can benefit operators, industry and, most importantly, consumers.
Given video is swiftly becoming a standard fixture in the consumer web experience, and live streaming of sporting events tipped to be the next tidal wave to hit mobile networks, any compromise in end-user Quality of Experience (QoE) is clearly unacceptable. As internet-connected devices and streaming services rise in popularity together in a virtuous cycle, each reinforcing the value of the other, the OTT industry growth will continue to accelerate. Faced with these facts, it’s up to the network operators to now make a strategic choice in terms of how they prepare their networks for the future of online video.