Cloud cover: The development of key applications

As operators recognise the limitations of traditional in-home service provision we are starting to see a steady rise in cloud technology adoption. Adrian Pennington looks at the development of three key cloud TV applications.

In order to match the heightened consumer expectation for TV everywhere, all service providers have cloud technologies on their roadmap. Rollout is hampered in many cases by existing investments in legacy on-premises equipment, copyright and legislative issues and some technical nuts that have yet to be cracked.

The benefits, though, of a shift to the cloud are widely understood. As John Carlucci, president and chief technology officer, Alticast, says: “What’s important about cloud is how it helps operators to break free from the bounds of the set-top box and use DevOps models and ubiquitous IT tools like HTML5 and JavaScript for app development.”

It’s also worth observing that cloud tech promises to revolutionise business models even further. “The next big step is providing events-based channels and disaster recovery capability and ultimately doing full linear playout of sophisticated channels in the cloud,” says James Gilbert, co-founder, Pixel Power. “In that light, cloud DVR and ad insertion are good tests of the infrastructure, but in terms of the future, the cloud is about much more.”

Cloud DVR

Some reflection of this shift in thinking is in the terminology which, in relation to DVR services, has shifted from RS-DVR to nDVR to cDVR. However, while cloud-based DVR is being deployed, “there’s no big bang shift in behaviour” according to Accedo’s EMEA vice-president Adam Nightingale.

Accedo expect over 50% of cDVR penetration to have occurred by 2022, roughly the timeframe that Ericsson also forecasts. Parks Associates research predicts cDVR subscribers worldwide will total 24 million by 2018.

Nokia, which recently took over Alcatel Lucent, says it has nine deployments for cDVR including Liberty Global, Telefónica, Numericable and Vodafone Portugal.

South Korea’s KT Skylife is among those using software specialist Alticast’s solutions to deliver either cDVR or hybrid systems that leverage the set-top box and cloud. Spanish cable service provider Euskaltel meanwhile launched cDVR last year using a Nagra solution to offer a multiscreen service on legacy set-top boxes and OTT to connected devices.

The upsides are numerous. For end users, content deployed in the cloud enables multiscreen access and recording of multiple simultaneous shows.  Without the limit of a 500GB hard disk, storage capacity is dynamically allocated. In addition, the relatively high failure rate of set-top disks and the consequent loss of personalised content can be mitigated. Customers can also retain recordings when they upgrade to new devices.

For the service provider there are capital savings from using shared and scaleable storage, and theoretically fewer service issue call outs. Ericsson says a typical truck roll costs US$75 (E67) per subscriber. This was the main driver for the deployment of cDVR by operators such as Cablevision in the US and KPN in the Netherlands, it suggests. “Associated costs quickly add up given that approximately 5% of hard drives fail per year and new features and upgrades require hardware replacement,” says Itai Tomer, head of cloud DVR, Ericsson. “A centralised network is more reliable than distributed drives.”

More importantly, cDVR delivers a greater degree of control over content for customer, service provider and advertiser alike.

“A content provider can assign rules around which content can be recorded and for what period,” says Roland Mestric, Nokia’s head of video marketing. “Such data can be used by advertisers who can place relevant ads around content when it is actually watched. You can also ensure the ads aren’t skipped.”

However, technical and licensing challenges are still hampering rapid advancement of this kind of service. Negotiated rights vary from programme to programme and country to country, with the pivotal issue being whether a unique copy is required for every subscriber. Each deployment has its own challenges and operators must ultimately decide between a private, shared, or hybrid architecture.

“In the US for instance, the landmark [2008] Cablevision decision deemed a unique copy saved per user as the standard, although we are still seeing MSOs looking to redress that decision independently,” says Tomer. “In Europe it varies according to region with issues ranging from high costs [of deploying a private copy solution], to securing the rights to content and consolidating within a variety of regions with different regulations.”

Matters are evolving, at least in some markets. For example, new legislation before the French Senate should provide a legal framework encouraging the deployment of cDVR solutions in France.

Private copy deployments are extremely costly. Tomer says, “A private copy system requires a unique copy of a programme to be saved for every subscriber that requests it, meaning recordings cannot be shared. Each single, unique copy of the programme has to be saved for each user, which requires a huge, growing volume of storage, very high recording and playout concurrency and that can be problematic to sustain.”

This plays into the three main technical challenges: scalability, flexibility and reliability. Regarding the first, says Mestric, “We need storage that can scale from a few TB to tens of petabytes. We need to ensure the platform can ramp from start to full deployment.”

More flexibility in the system is vital to make sure the service is optimised based on service or use case requirements. “With cDVR you can deploy a number of end user services like live restart or catch-up,” says Mestric. “The characteristics of a system for live are, however, not the same as for VoD or catch-up, so based on the requirements of operators a system has to have the flexibility to be optimised. Moreover, if an operator starts with catch-up they need to be able to add full cDVR as it negotiates rights with content providers.” Platform robustness ensures consumers have access to content and don’t lose it. “If a hard disk fails it needs to be recovered very quickly. Performance and speed is the key criteria,” he says.

At the network level the technology is now available from streaming and content delivery network vendors such as Harmonic or Edgeware, enabling robust solutions to be deployed. Of course, ‘cloud’ in cDVR doesn’t mean a ‘public’ cloud. “Operators are, for the most part, still skittish about public cloud due to concerns about security, quality of service, and control. So these cDVR deployments are happening on private clouds,” says Yuval Fisher, CTO, MVPD, Imagine Communications. “The general view is that a cloud is an infinitely scalable and fungible collection of resources. But the reality for cDVR is that this doesn’t scale. As a result, cDVR deployments require specialised clouds, and this is something the industry is just now digesting.”

The verdict on cDVR is that it confers benefits to operators and users by enabling the delivery of content to multiple screens. “Rights holders are also starting to realise that cDVR actually offers new opportunities to monetise content with services that appeal to a new generation of multiscreen TV viewers that value convenience and flexibility more than any other market segment before,” says Simon Trudelle, senior director, product marketing, Nagra.

However, says Tomer, “cDVR isn’t a trivial deployment and legal issues, storage concerns and performance requirements must be considered.”

Cloud ad insertion

Ad insertion has long been seen as another central application of cloud technology. However, generally speaking, operators are yet to fully embrace the cloud to deliver ad insertion across live and on-demand services.

According to Thinkbox, citing 2015 figures, linear still accounts for 81% of all TV broadcaster viewing in an advanced market like the UK.  As overall cloud-based on-demand TV consumption increases, the value in managing addressable ads delivered to personalised, connected screens will become more transparent to broadcasters, brand advertisers, and measurement firms.

“The deployment of next generation fully connected set-top boxes, complemented by other screens, is clearly the enabler that will drive demand,” says Trudelle.

While early adopters are beginning to implement ad insertion technologies, others are in wait and see mode. “It’s early days,” says Tomer. “One thing is clear: operators agree that changing viewing habits combined with OTT video and innovation in cDVR technology have changed the game for advertising.”

The promise of cloud-based delivery in relation to advertising is that it can help enable a much more effective and deeper degree of targeting than has hitherto been possible.

“The real value of the cloud here is targeting,” says Imagine’s Fisher. “A different way to look at this is that as advertising moves towards impression-based targeting, it is also moving toward software and cloud deployment as a natural evolution in how solutions are packaged.”

Cloud enables scaling – the ability to utilise more ad insertion resources when needed, such as at prime time. This is “basically a cost saving feature” for Fisher. “So it’s not really a benefit as much as a mitigation of the extra cost associated with deploying full targeting capability. More important, new ad insertion mechanisms are based on software, and that brings significant operational simplicity.”

Cloud ad insertion also overcomes key problems with ad blockers, so operators can increase revenue from ads, although it won’t totally eradicate blocking.

“Generally the user experience is better with cloud ad insertions but probably the most important advantage is that an operator is in control of the ads, rather than having to rely on the platform owner,” says Accedo’s Nightingale.

Whether server-side or client side, ad insertion is now well defined from a technology standpoint. Client side solutions include using one player to play the main content while a second is used to play ads. Nightingale says this solution “permits the app to initialise the players ahead of playing the content and then switches players [brings a player to the foreground] when required.” He says the customer doesn’t see any buffering or delays between the main content and ads. “This solution generally works well but it does require a lot more memory on the client device.”

Alternatively, client-stitching applications make use of one player, and while playing either the ad or the main content, are able to buffer the other content for the next item ready to be played when required. “No buffering is seen by the customer and the UX is seamless,” he says. The player can also be given a playlist of assets that play sequentially one after another.

Dynamic stream stitching, sometimes called manifest manipulation, is performed on the server side and requires very little customisation of the client side player.

Nightingale says: “The client tells the server what content the customer wants to play and any ad requirements and the server makes the calls to the ad server. The ads are stitched into the main content and delivered to the client.”

While cloud ad insertion requires greater pre-processing of ad content in terms of the use of adaptive bit-rate, video quality, sound codecs and so on, the challenge service providers face is often more on the business front.

“With the exception of the US market, where MVPDs actively manage some of the advertising space on behalf of broadcast networks, demand for addressable ad insertion remains low as the ad space is managed by broadcasters,” says Trudelle.

For scenarios where set-top box connectivity and interactivity are not guaranteed or are limited to a subset of the subscriber base, the cloud may not achieve value in the short-term.

“It really means that service providers should have a plan to go ‘fully connected’ before envisioning deploying cloud ad insertion,” says Tomer.  “Those who can move fast will clearly reap the benefits of this new technology.”

Cloud TV UX

Transplanting user experiences to the cloud offer many of the same advantages to operators, notably the ability to change the UX rapidly and at scale, rather than rewriting UXs for every make and model of set-top box, and enabling an operator to tailor personalised UIs.

“The always-on nature of the cable network enables the cloud to be harnessed so that operators can deliver advanced user experiences – such as Millennial navigation, kids’ modes and sports zones – that are not capable of being supported by the set-top box alone,” says Alticast’s Carlucci. “Since UXs no longer need to be resident on the set-top box, operators – and customers – can have an infinite number of UX variations.”

With Accedo AppGrid, users can update applications in real-time across platforms and control it per user, country or time of day without redeploying and submitting applications. “Being cloud-based means providers can easily engage with viewers, for example through in-app notifications,” says Nightingale.

UXs delivered as streams to every set-top box enable operators efficiently to deploy services “that are equal to – or better than – experiences that run on the box itself,” says Murali Nemani, CMO, product management and marketing, ActiveVideo.

There are a number of instances illustrating how Cloud UIs can bring a diversity of advanced UXs to existing set-top boxes: VoD and catch-up services with Ziggo in the Netherlands; trend-driven UIs with multiple tiles of live video on single tuner set-tops with Liberty Puerto Rico; and the complete YouTube experience to upwards of 500,000 existing STBs at UPC Hungary – all of which are ActiveVideo deployments. Similar innovation is taking place in the US, where Cablevision was the first to make the full Hulu experience available on all its current-generation boxes.

Cloud-based and set-top strategies are not mutually exclusive. “You’ll see the industry continue to use the cloud to deliver TV UXs even as boxes become more capable,” says Carlucci. “We will leverage what the improved STB can do but we also will continue to see the cloud and the network evolve.”

There’s a similar strategy at ActiveVideo which offers GuideCast and StreamCast products to ensure that next-generation services can be delivered at scale to set-top boxes already in customer homes, as well as to new devices coming to market.

“Some [operators] are streaming the entire UX from the cloud; some are using cloud resources to complement a next-gen device strategy,” says Nemani.

US cable company Charter has moved to stream its entire UX from the cloud beginning with its Spectrum Guide. Tom Rutledge, Charter president and CEO, speaking at CES last year, said: “We’re taking the intelligence out of the box and putting it into the network and making the box a thin client box so that the processing power of the box is no longer a relevant issue; the processing power moves to the network. That’s a breakthrough.”

Craig Moffett, partner and senior analyst at MoffettNathanson Research, estimates that Charter will spend US$2 billion (e1.8 billion) over five years for its cloud-based guide build compared to US$8.4 billion for IP-enabled boxes – a 77% capex saving.

However, some industry participants express caution. While agreeing that cloud UXs can tap “virtually unlimited back-office CPU power” Nagra believes this is “driving the industry to a position of compromise” because the functionality of a native UI/UX “cannot be replicated with today’s cloud UI offering.”

Anthony Smith-Chaigneau, senior director, product marketing at Nagra, says: “Cloud UX deployment has its share of technology challenges. These include the latency of the remote control, as each action of the remote has to be transmitted to the cloud for processing. If network resources are limited it is difficult to anticipate the actual load of the network. This is the case in particular for live/linear services where each video stream is unicast.” There’s a real question about the simplicity of the set-top/CPE client, he suggests. “Both video and audio still need to be decoded, taking into account the numerous compression and transport formats, this requires a variety of computing power requirements.” Smith-Chaigneau also asks how providers like Netflix or YouTube will react to being “proxied” by a cloud infrastructure. Currently they have their UI implemented in the client device.

“Ironically TV everywhere is addressing laptops, smartphones and tablets that have enormous computing power,” he says. “So with a cloud UI-UX are we just talking about the issue of ‘incapable’ set-top boxes/CPEs in the field? Is that the problem that we are addressing with this solution?” He answers his own question: “It may well be that the cloud UX is the solution for small and medium operators that want to deploy advanced services and an advanced UX without having to bear the cost of implementing a middleware in the client set-top box/CPE. It could at least support a middleware that provides mainly the video and audio rendering without a PVR, video gateway or home network. Network bandwidth still remains a challenge, but there might be fewer problems as these operators have to serve smaller number of clients.”