opportunities to leverage the IP technology for monetizing providing new value added
services to legacy cable consumers. This article is based on a white paper by Farncombe
sponsored by NAGRA.
The rise of over-the-top (OTT) video distribution raises multiple challenges for cable companies – strategic as well as technical. On the strategy front, they are facing the prospect of cable-like entities emerging ‘in the cloud’, able to bypass their managed cable networks to sell content directly to consumers, using next-generation networks (including wireless), new standards and new technologies to guarantee QoE – all of which may tempt their customers to ‘cut the cord’. Meanwhile, not only are cable operators’ premium offers being increasingly undermined by content illegally distributed over the Internet, but they are also being eroded by some of the very measures being undertaken to counter-act online piracy – for example, the Hollywood studios’ Ultraviolet model, which grants streaming rights that overlap with the existing VOD window.
If working out a strategic response to these complex developments is fraught with difficulties, so too is choosing the optimal array of technology options.
Ideally, cable operators want to be able to offer just as much content choice at any time, on just as many devices, and to just as many locations, as their new OTT-based challengers – and to offer it securely. But how to deliver this?
Adding PVR functionality to STBs, and then on-demand movies via QAM and DOCSIS, was once a challenge, but is now relatively easy; it is far more difficult to deliver IP video to multiple (often non-interoperable) devices with different DRMs. Complex choices are involved: for instance, should delivery be cloud-based, from the headend, or through an in-home gateway (or even a combination of all three)? Upgrading the business and back-office end of the network is only the beginning: how to deal with all those legacy devices with no OTT stack at the other end? How, indeed, is a cable operator to address all those different devices it no longer ‘owns’ and rents out?
In this article we seek to explain how new standards and technologies can be leveraged to create a flexible multiscreen model for cable operators, with the objective of maintaining competition and garnering fresh revenue streams, while – hopefully – exerting downward pressure on subscriber churn.
We also argue that there are many benefits to be obtained by migrating to an ‘all-IP’ video distribution model, particularly in view of the backend efficiencies to be obtained, as well as simplified bandwidth resource management and set-top box integration.
Although an ‘all-IP’ model should be the ultimate solution for cable operators both on-net and off-net, we acknowledge that their investment in a DVB-C/QAM-based network cannot lightly be written off: a phased approach, beginning with the conversion of on-demand services and some niche linear services to an IP-based adaptive bitrate (ABR) approach, is likely to be most realistic in the foreseeable future.
We believe this will deliver multiple benefits to cable operators, not least in being able to use common technology to reach all target devices and to extend services ‘off-net’ to potential new subscribers outside their existing cable coverage areas.
Multiscreen – the pay-TV battle-ground
It all started with the PC: as an entertainment device it was initially confined to playing back music and movies via optical disks; but the arrival of ubiquitous highspeed broadband and user-friendly Web-based streaming portals transformed it into a viable alternative to the main TV display in the living-room. This was the beginning of the ‘second screen’ revolution.
This was followed by two separate, but parallel developments:
1. First, connected games consoles brought the PC-based video streaming experience back to the TV set, a trend taken up first by a new generation of broadband-enabled digital TV set-top boxes, and subsequently by the evolution of the ‘smart’ or connected TV.
2. Second, the iPhone, and then the iPad (together with all their imitators) built on the PC-led ‘second screen’ phenomenon, turning it into a massmarket proposition.
The paradox is that viewing of linear TV remains surprisingly buoyant despite this, showing online consumption to be additive rather than substitutive. In the same international study, Ofcom found that minutes of linear viewing per head per day have been increasing in the overwhelming majority of countries surveyed.
However, it is also true that much of this scheduled linear viewing is taking place alongside consumption of content on ‘companion’ second-screen devices.
Google’s latest research from the US, in partnership with Sterling Brands and IPSOS, shows that for 77% of the time people are using a TV, they are also using another device simultaneously – 49% a smartphone, 34% a PC/laptop.
Cable operator takeaway: Linear viewing over the managed QAM network continues to deliver value – but is increasingly shared with unmanaged devices
The over-arching implications for cable operators are that:
a) their subscribers want to view content (whether operators’ or someone else’s) at a time and place of their choosing, on any device they happen to possess; and
b) if operators fail to make content available in this way, their customers, legally or otherwise, will seek to consume it on devices they do not control.
For these reasons alone, this makes migrating to a multiscreen approach an attractive option for cable operators. But there are also sound operational reasons for doing so, since, in effect, ‘going multiscreen’ is synonymous with transitioning to IP. The business implication is that migrating to IP can deliver operational flexibility that may drive capex and opex savings, because:
• IP set-top boxes based on web browser technologies are less complex than traditional legacy cable ones, the product cycle is shorter, and they therefore have a lower cost of ownership
• Managing an end-to-end IP infrastructure provides enhanced flexibility, and for cable would result in synergies with the broadband delivery platform This means that for cable operators, going multiscreen is not just one strategic option amongst many, but a natural evolution that will soon become a ‘must-have’.
The evidence is that cable operators are increasingly addressing this business imperative. In a paper on multiscreen gateways, Cisco quotes a 2010 global study by Heavy Reading that found that 40% of cable operators were planning to launch some type of ‘IP Video’ service by 2012 – with nearly 80% saying they had already begun or intended to begin ‘IP video’ trials, pilots or rollouts.
Cable operator takeaway: Multiscreen is a natural evolution that will shortly become a must-have.
Multiple devices using multiple technologies
Most cable operators recognize that they have to evolve towards the creation of a multiscreen ecosystem for their customers, if not a multiplatform one. But which of the many new technologies available should they use, and in which combination?
For a cable operator, a multiscreen strategy may be defined as one which is designed to reach devices unable to access linear or on-demand services via their DVB-C/QAM based cable infrastructure.
Generally, aside from PCs, laptops and games consoles, we are mainly talking about iOS and Android devices here. Two issues arise:
• first, these devices are likely to be controlled by someone else (e.g. Microsoft, Apple or Google), which raises issues about the business costs involved in reaching them
• second, they are using multiple flavours of Internet technologies such as adaptive bit-rate streaming (ABR) technologies and content protection schemes.
This situation is slowly changing through the creation of a new generation of ‘open’ standards and technologies that have their origins in the Internet and mobile sectors and are biased towards open-source models – examples include DNLA, MPEG-DASH, HTML-5, HbbTV, and a variety of IP protocols.
However, while cable operators should not ignore these developments, in practice the target devices’ differing operating systems imply the use of different software stacks and UI solutions today, as well as the deployment of a system which manages multiple, non-interoperable DRMs in an IP-based ABR environment. We examine both ABR and DRM technologies in more detail in the next section.
Cable operator takeaway: Multiscreen implies the use of ABR technologies and the ability to manage multiple DRMs
Adaptive bit-rate streaming
One of the major obstacles to take-up of IP video delivered to second screens over un-managed networks has been the inability to ensure smooth delivery of highquality video.
Adaptive bit-rate (‘ABR’) streaming addresses this by reacting on-the-fly to sudden decreases in available bandwidth by lowering the quality (through decreasing the resolution, frame-rate or increasing the compression level). This is effected by calling up a lower bit-rate stream until the constraint has disappeared, helping to side-step issues such as screen ‘freezes’ while the client waits to receive extra videobits (‘re-buffers’). (See Figure 3).
This may degrade the picture, but practical experience suggests viewers much prefer temporarily lower-quality pictures to a complete interruption to their linear viewing experience.
An example of a new open standard using ABR, which emanates from the mobile sector, is MPEG-DASH. Ratified at the end of last year, it supports delivery of video content over existing IP and mobile networks – with the aim of becoming the global standard for ABR streaming. MPEG-DASH has been adopted for use in France’s next-generation hybrid DTT platform, TNT 2.0; for the UK’s hybrid platform Freesat; and by Abertis Telecom in Spain for its OTT service.
In practice, the general increase in consumer broadband bandwidth around the world has meant that ABR streaming is now preferred to download technologies, which were originally adopted to preserve video quality and uninterrupted consumption by allowing storage and replay at local level for some or part of the viewing-time.
Another advantage is that ABR also allows advanced services to be offered, such as ‘trick-play’ features (for instance, pausing or rewinding a movie). As we shall see in a later section, this distinguishes it from video streamed via IP Multicast, where – for instance – if a user pauses the video, the stream has to be switched to unicast mode.
New DRM approaches
In the past, proprietary bundling of the browser, the media player, the content encryption and the DRM have implied severe fragmentation of DRM support across different CE devices. As a result, service providers have had to duplicate their content in multiple formats to target all devices.
A key feature of standards such as MPEG-DASH and HbbTV has been the introduction of common encryption – allowing the use of multiple DRMs with the same encrypted ABR asset. This is also a notable feature of the new Ultraviolet electronic sell-through and streaming platform promoted by the Hollywood studios through the DECE industry grouping.
The aim is to support the monetisation of hybrid platforms through streamed premium content, a feature which has so far been lacking from the connected TV business model.
This represents a radical change to the traditional vertical pay-TV set-up, where only one content security solution was generally supported by an operator’s box, and signals an industry-wide attempt to address the issue of interoperability for OTT video. In connected homes containing multiple second screens of different types, different content security solutions represent a barrier to take-up.
This evolving model borrows from the concepts behind the DVB Simulcrypt scheme, in which content is created in a common format and/or encrypted through a common encryption scheme, with a new DRM layered over the top to enable decryption on previously non-interoperable devices.
Cable operator takeaway: New Internet/mobile-based technologies and standards mean that it is becoming easier for cable operators to extend their operations into the OTT space. ABR and interoperable multiple DRM solutions are particularly useful.
The ‘all-IP’ cable model
An ‘all-IP model’ means different things to different people. For some, it entails pursuing an IP Multicast model from end-to-end – i.e. not just between the headend and local hubs, but all the way down to the consumer. For others, it means an IP-based, unicast, ABR approach.
If multiscreen is a cable operator must-have, we have argued that it requires the use of IP-delivered video using ABR in unicast mode – because this is what the most rapidly- growing consumer devices (e.g. smartphones and tablets) use. IP Multicast is not (at least today) a technology that will run across the open Internet. Moreover, few consumer devices, let alone smartphones & tablets, support it.
ABR has a number of other advantages: in a digital TV environment, switching channels can often be slow. ABR can, however, soften the blow by quickly loading a low bit-rate stream first, and then moving on to higher-rate streams as the bandwidth becomes available.
The technique is also based on the tried-and-tested TCP protocol, which has error recovery already built into it: if any video packets are lost because of a bandwidth squeeze, they are automatically re-transmitted.
This contrasts with what happens with IP Multicast, which is based on much less robust protocols: when traffic congestion occurs, it is more likely to lose video packets and never recover them.
Meanwhile, as mentioned earlier, the network capacity benefits of adopting IP Multicast rather than an IP unicast approach are often over-stated. For instance, when using ABR to deliver video via a CDN which implements effective edge-caching, there is no impact on network capacity when additional streams are requested by subscribers. IP Multicast also encounters many problems in an environment where there are mixed downstream bonding groups (DBGs). This is an issue for cable operators rolling out IP Video, since the number of downstream channels in a DBG is likely to increase over time as home gateways become progressively more advanced. This results in an array of different gateways in the field, accessing different DBGs.
Using ABR and unicast delivery, however, every gateway modem can exist in a bonding group suited to its capabilities, allowing multiple bonding groups to overlap, and thereby increasing the efficiency of bandwidth usage.
For these and other reasons, the authors of a recent paper by Motorola Mobility, John Ulm and John Holobinko, conclude that “under most conditions, multicast delivery will have little or no Source: bandwidth capacity advantages over ABR unicast delivery.” Figure 6 below shows some results from that paper, noting that for early IP Video deployments, penetration rate will be low, and that there is no multicast benefit below a certain level.
It might appear that the logic of our position is, therefore, that all the QAM broadcast channels should be converted to DOCSIS, with the entire offer, linear as well as on-demand, pushed out to customers as IP Video using an OTT/ABR delivery approach.
While we believe this should certainly represent the ultimate goal for cable operators, in practical terms this is very challenging.
The investment in a DVB-C, QAM-based infrastructure represents an enormous sunk cost that cannot be casually written off in the short-term. Meanwhile, DVB-C/QAM remains a very efficient technology for delivering broadcast channels with a guaranteed QoS, particularly HD – and as we have seen, linear, scheduled viewing hours remain buoyant the world over.
So what we are advocating is an approach which assumes that the ultimate goal is an all-IP approach using ABR streaming in unicast mode to deliver a full multiscreen eco-system, but pragmatically recognizes that this will need to be broken down into four stages:
• Step 1: Implement an IP-based ABR unicast approach for new services, which incentivises subscribers to use their managed network and services (for example, beginning with the creation of a new content-preparation chain able to target iOS and Android clients)
• Step 2: Move any QAM-based VOD services to OTT and unify back-end management through a single IP Video platform.
• Step 3: Migrate niche broadcast linear services (those channels with on average less than one viewer per node) to IP-based ABR.
• Step 4: Migration of all QAM video to IP using ABR.
Clearly, for most cable operators, Step 4 is quite some years away. It also requires the DOCSIS network to be scaled for concurrent usage of linear TV during prime time.
Thus in practice we are advocating Steps 1, 2 and 3 in the short-to-medium term. Naturally, for cable operators planning their next-generation networks, the results of the cost and capacity trade-offs involved in each of these steps are exceedingly complex and will vary depending on the current state of their network and their perceived customer needs.
NAGRA accordingly recognizes that the process cannot be rushed: "While the shift to multiscreen means the ultimate goal of cable operators should be all-IP delivery," says Hitesh Lokhandwala, NAGRA General Manager India, "this will be difficult to achieve in the medium term and will take some time to accomplish given the network upgrades and set-top box replacement required. We foresee a long period where higher viewership broadcast service continue to be delivered to the home via DVB-C alongside a broad range of linear and on-demand IP services."
Cable operator takeaway: While the ultimate goal of cable operators should be all-IP delivery, this will be difficult to achieve and take a long time to accomplish given the network upgrades and set-top box replacements required. But there are many beneficial steps that can be taken towards this goal in the meantime.
for more information on "Farncombe white paper sponsored by NAGRA" please visit http://www.farncombe.com/ registration-wp4-china/