The Streaming Video Alliance (SVA) was boosted in May 2018 by Google joining.
The Streaming Video Alliance (SVA) has been urging online service providers and vendors of technology for delivery of video over the Internet to rally behind common standards.
The SVA’s Executive Director Jason Thibeault has consistently argued that the streaming industry must move away from the fragmented, complex and in some cases antagonistic proprietary video architectures prevailing today to gain the scale required for mass deployment of OTT services. Without such common standards the march from traditional broadcasting to OTT will be slowed down as consumers become frustrated by their new viewing experiences, while broadband network operators will see their infrastructures bogged down by all the additional traffic. There has been some evidence of this during the current FIFA 2018 football World Cup, with a number of broadcasters and operators experiencing the glitches such as buffering and visual artefacts that they hoped had been confined to OTT history. In some cases, their networks had been overwhelmed by traffic and unable to sustain normal quality of service.
The SVA was set up in 2014 as an umbrella organization defining the broad framework at a high level for OTT video services based on interoperable infrastructure components. Other longer standing standards bodies are then developing the detailed technical specifications, according to its Executive Director Jason Thibeault. Such bodies include MPEG with its DASH adaptive bit rate streaming technology, the Alliance for Open Media on codecs optimized for online environments and the World Wide Web Consortium (W3C) with the HTML5 markup language incorporating support for video.
The SVA aims to assimilate these standards into a common higher-level framework that will enhance scalability and improve performance, with latency kept to a minimum. Thibeault stressed the SVA would keep working with these bodies and other players in the video ecosystem to ensure that the standards would coexist and that any gaps were plugged.
This distinguishes the SVA from some other umbrella standards bodies such as the Ultra HD Forum that do not define standards themselves at all. The SVA will step up and develop specifications where they are lacking, as it has been doing over the three related areas of Open Caching, live streaming and Quality of Quality of Experience (QoE) measurement. These are being developed by SVA working groups that are collaborating to ensure they are aligned and dovetail properly. One point here is that while there are some effective proprietary technologies designed to minimize latency over live streaming services, such as Net Insight’s Sye, there are no universally agreed mechanisms.
The Streaming Video Alliance is focusing on high level standards, according to its Executive Director Jason Thibeault.
The Open Caching Working Group was set up firstly to identify core components of a caching system and secondly to define the basic architectural guidelines to implement these. The underlying aim is the same as providers of CDN (Content Deliver Network) services have always had, which is to stage popular video content from inside the operator network to locations close to consumers. Currently CDNs achieve this at a lower level through proprietary mechanisms with no universal cache function, which is what the SVA is developing to underpin standards based interoperable global OTT video services.
This open caching system will operate by allowing existing CDNs, irrespective of the provider, to continue using their own caches in combination but then combine them with distributed open caches based on the SVA standard. This will involve a hierarchy of multiple delivery tiers organized so as to optimize the balance between storage, compute resources and bandwidth.
Within this hierarchy will be three core component sets. One will include the end points or open cache nodes (OCNs), which deliver content to users within the service provider’s domain and at the same time capture logging and billing data for transmission back upstream to the source CDNs or services. Then the second component type are service provider open cache controllers, which interwork directly with the CDNs via APIs defined by the SVA. A typical large operator would probably have thousands of such OCNs clustered around a smaller number of the open cache controllers.
Then the third set of components are at the CDN end. These are the CDN Open Cache controllers, which are distinct from the service provider open cache controllers and communicate with the multiple operators whose traffic is transported over that CDN.
The Open Cache complements another SVA task area associated with reducing latency, concerned directly with live streaming. Although this is covered by another working group, it overlaps with the open cache work because that will also help reduce latency by staging live content efficiently to caches during its delivery. The live streaming work is focused directly on the whole latency path between camera capture and playback on the user’s device. An important aim, as Thibeault made clear, is to synchronize OTT with broadcast services as is important for some second screen applications, as well as the overall user experience.
Both the open caching and live streaming work will rely heavily on effective measurement of the user experience if they are to meet their objectives of closing the quality gap on broadcasting. This is why the respective SVA working groups are liaising with a third group dedicated to QoE measurement.
This QoE Working Group began by defining four key metrics as effective surrogates of the experience, firstly video startup measured in seconds from initiation of play until the first frame is delivered. Second is rebuffering ratio, which is time spent rebuffering divided by the total time taken from beginning till end of playback. Third is average bit rate in bps and fourthly whether the video started correctly without timing out. The SVA then has to define how to measure or calculate these relatively simple metrics at the device level and then bring them back to the operator. Analyzing the measurements and acting to restore quality if it has dropped is then up to the service provider using the tools the SVA has provided.
One side effect of the SVA’ focus on open standards is that some of the major CDN and service providers have proved reluctant to support the work because they want to extract advantage from their proprietary ecosystems. So while Google recently joined Amazon Web services (AWS), Cisco, Comcast, Intel, Liberty Global, NBCUniversal, Nokia and Verizon as SVA members, Apple, Microsoft, Akamai and Netflix have yet to join. The SVA is actively lobbying these big players to join according to Thibeault, because it is aware it does not yet represent the whole streaming industry. However, it has sufficient momentum and enough big hitters on board to ensure that its standards are still likely to enjoy wide implementation.
You might also like...
Most CDNs (Content Delivery Networks) and the video streaming services they support have coped well with the peak time loads of the current FIFA World Cup, beyond a few well publicized glitches.
The Cloud itself is pretty simple, but for broadcasters, leveraging it can be complex. It isn’t that difficult technically, but Broadcast still straddles requirements that demand real time, high resolution, quick turnaround solutions; and this is coupled with the n…
While it’s clear that High Efficiency Video Coding (HEVC) is superior in compression performance to its AVC predecessor codec format, getting the essential patent holders of the standard to agree on royalty terms for use has proved to be f…
It is not just broadcasters and pay TV operators that have struggled to cope with the accelerating momentum behind OTT, because it has been just as challenging for their technology providers.
In case you missed a day with The Broadcast Bridge, here are two popular articles that may be lost in your inbox. The first of two articles looks first at what seven manufacturers think about cloud solutions and second examines…