RDK: the progress report

This year's RDK conference in February

This year’s RDK conference in February

The RDK software framework has won strong backing – not least from Liberty Global in Europe. With new extensions tackling fresh categories of device, can it fulfil its promise of bringing web-speed innovation to cable? Stuart Thomson reports.

Web-speed is the new watchword for traditional pay TV providers. Much of the innovation driving the evolution of the TV experience is now coming from OTT players and app developers. Meanwhile, the spectre of an emerging generation of ‘cord-nevers’ – younger consumers who have never signed up for a pay TV service and see no reason why they should – has unsettled companies that no longer hold a monopoly on video distribution, even to the big screen in the living room.

The problem of OTT upstarts being able to provide a more compelling user experience than cable and other service providers, combined with the flexibility that web distribution gave them to experiment with different designs and functions, has given rise to efforts by slower-moving pay TV operators – by the cable industry in particular – to cooperate in the face of the new threat.

The Reference Design Kit (RDK) was cable’s answer to the problem of being outflanked by OTT providers in delivering a compelling TV experience to users. Conceived initially by Comcast and taken up by other US cable players, the RDK adopted the open-source software model that has underpinned much web innovation and tried to adapt it to the cable world. Industry players including vendors and operators collaborate by contributing code to the project that is shared amongst licensees.

The RDK software stack is designed to create a common framework that will enable a degree of interoperability for applications between devices. The stack, unlike previous attempts to create a common middleware – such as the MHP – is not a standard as such and does not go beyond a common baseline – namely the point where operators can differentiate their offerings through the creation of revenue-earning added-value services. This is left to them to do on top of the RDK on a proprietary basis.

Growing interest

RDK software is now believed to be deployed in over five million devices with multiple operators in multiple markets.

According to Steve Heeb, president and general manager of RDK Management, the independent body in charge of administering the RDK, a number of European operators are at different stages of deployment of the platform.

Most vendors involved in RDK also say they have seen significant growth in interest in the platform over the past few months as operators look to implement new TV features in order to compete with the likes of Netflix and other OTT sources of video. Interest is not confined to the cable operators that originally backed the concept but also extends to IPTV providers, according to Jaison Dolvane, CEO of TV software specialist and RDK licensee, Espial.

“We certainly have seen significant interest that has increased over the last six months. We ourselves have secured tier one customers in North America and Europe. It’s not just cable – telcos and satellite players have expressed interest,” says Dolvane. “The problem around legacy technology is more amplified in cable but they and the telcos have shared problems innovating. They still think about [implementing changes] over 12-18 month periods, when competitors like YouTube and Netflix can change quickly. Part of the reason why operators have this type of problem is the legacy of closed systems with middleware that has blocked them from doing too many new things.”

Dolvane says that the RDK can be adopted by operators irrespective of their scale, but the means by which they approach the implementation may vary. He suggests that operators with a million-plus subscribers will look at RDK seriously. Smaller providers will likely require an off-the-shelf product to deliver advanced TV. This could of course be RDK-based, with vendors providing a pre-packaged application framework on top. “As we get more operators adopting it we will see more off-the-shelf solutions,” says Dolvane.

Liberty Global

Adoption of RDK has been led by US operators including Comcast, the originator of the project. However, the principal European backer of RDK to date has been Liberty Global, which has used the RDK framework to launch the latest iteration of its Horizon advanced TV platform for UPC Poland, previous versions of Horizon in western Europe having been based on a proprietary platform.

Peter Hahn, director of product management at SeaChange, says that the Polish deployment “proved that the flexibility of RDK would work in commercial deployments outside the US”.

The deployment went live in January after significant delays, due to a variety of factors including the complexity of integrating DVB conditional access, the hooks for which had not yet been defined as part of the software. Other aspects that required detailed work included picture-in-picture functionality and getting channel change right. SeaChange contributed code to RDK to provide the required extensions to support DVB features.

Liberty Global’s implementation of RDK for Horizon in Poland left SeaChange with plenty of room to deliver innovative software on top, according to Hahn.

“RDK is an important part of the software stack but it is not the complete picture,” he says, pointing out that the user experience makes use of SeaChange’s own Nucleus software platform to deliver the end-user experience by providing the business logic and application frameworks. “This is above the RDK layer – all that needed to be optimised to meet Liberty Global’s requirements,” he says.

Hahn says that Liberty Global now has a platform in place that will enable it to create a much faster user interface development cycle because the UI is independent of the underlying software in the box. “You can change things on the UI layer without changing anything else,” he says.

Alternatively, the same UI can easily be ported to different RDK-based set-top boxes. In the case of Liberty Global, one result should be that the operator can deploy Horizon in different markets much more easily than before.

The RDK is designed to enable operators to reduce costs and increase the speed of new service delivery, but it requires proprietary software on top to deliver a meaningful service. “One benefit that our product provides is that it can run in different configurations. We have in-home distribution and our stack can run over a headless gateway or over a tiny USB stick with the same software spec or OTT back office,” says Hahn. “We can distribute the signal in the home through MoCA or WiFi or Ethernet or whatever is supported in the box.”

While Hahn says SeaChange weighs up “every feature and line of code” to see whether it could better be rendered in the cloud than in a set-top, he adds that operators  “still see the need to deploy boxes to control the user experience and have a presence in the living room”.

“They want to have a spot on the table where the remote control is,” says Hahn. “But they also want to limit the time to market and be more agile in rolling out new features. This is where you see a trend of moving features like the UI into the cloud.”

The advantages of cloud-based provision of things like DVR management, schedule management and conflict management need to be weighed against the challenges that cloud-based delivery can face in the real world through bandwidth limitations. The RDK itself is neutral, says Hahn, because it just provides a set of standard consumer premises device features.


With European deployment in mind, the RDK recently embraced extensions to cover the requirements of European networks, including DVB extensions covering things like Teletext and subtitling and, more recently, support for DVB Service Information – the elements that bind DVB transport streams together and allow EPG data to be transmitted. These have minority use cases; many European operators are moving towards or have already moved towards sending EPG and other information out-of-band via IP rather than as part of the DVB transport stream, but others prefer to continue with in-band transmission.

Further likely developments include support for IP multicasting, seen as a crucial element for future video delivery.

Heeb says that DVB conditional access components added last year are now integrated with the RDK trunk. DVB SI components, for which Arris contributed code, became available in February.

“In the case of DVB SI there were multiple contributors. Sometimes a lot of people contribute and then they see which is the best to adopt. They select one vendor submission and work with that. The onus now falls on us to fix the bugs,” says Charles Cheevers chief technology officer, customer premises equipment at Arris. Cheevers says that, while DVB SI may only be used by a minority, European operators in general are committed to supporting QAM transmission of linear channels into the future.

On the conditional access side, the two leading DVB CA providers have already plugged into the RDK, and the third-ranked player is expected to do so soon.

The RDK does not currently support the European CI+ common interface standard for conditional access modules, but Cheevers says a number of operators are looking for ways to support this and the HbbTV hybrid broadband-broadcast interactive TV standard.

Heeb meanwhile says that extensions to the trunk to support IP multicast and multiple DRMs are likely in the future.

Despite all these signs of recent progress, RDK is not without its critics. One company notable by its absence from the list of RDK licensees to date is Nagra, which has focused on developing its own full middleware stack – OpenTV 5 – instead. According to Anthony Smith-Chaigneau, senior director of product marketing at Nagra, the regional and network differences between US cable and DVB cable networks mean that implementations will in any case require adaptation, “which means time, investment and testing as well as actual large scale deployments required to prove that any software can prevail”.

Smith-Chaigneau contrasts RDK, which he describes as a “project-oriented integration”, with OpenTV 5, which he characterises as a “product”. He adds that OpenTV 5 itself uses many of the open-source web technologies that RDK has embraced, including HTML5, web-page rendering framework Webkit and GStreamer, the open-source media player framework, “so we believe that we have already proven that we can deliver a full CPE stack in accordance with the cable operators’ vision”.

“The disadvantage of the RDK approach is that it is only a very small piece of an extremely complex puzzle.  You have to deliver the UI-user experience, application framework, broadcast technologies, headend and cloud technologies that all make up a modern cable television offering and that will entail quite some bespoke work on a per deployment basis,” says Smith Chaigneau. “There is a lot more to define, build, control, support and maintain. You cannot, with a small baseline software package for a CPE, deliver all these services – that costs time, investment and money and ownership for maintenance and support.”

Room for innovation

On the other hand, supporters of the RDK say the structure and self-imposed limits of the framework has worked well.

“We structured this as an open-source licence, and for companies used to the open source world this is not an issue. For others it is an education process,” says RDK Management’s Heeb. “Licensees get full access to the RDK source code and continued access to the broader RDK community – it is an education process for people not used to working in the open source world.”

John Maguire, director of strategy and marketing for TV technology specialist S3 Group, which is responsible for managing the RDK on behalf of RDK Management, says that the RDK now has about 200 licensees and that activity on the RDK portal is significant and growing. For Maguire, the division between what the RDK covers and what is left to vendors or operators to decide on is overt.

“The position today is very clear [and] there is clear strategic direction from the top level of RDK Management that keeps everything on the straight and narrow,” says Maguire. RDK’s focus on “the basic plumbing” is clear and extensions for different regions are commendable, “but differentiators that are service-specific belong above the line”, he says.

The RDK’s supporters say the framework is not designed to provide a complete solution to enable operators to deliver a service, but rather to provide a base on which they can then build their own differentiated offerings. “It provides part of the solution but it requires system integration and apps to complete it,” says Espial’s Dolvane.

The RDK approach essentially envisions three layers – the core platform, RDK extensions and the application layer, where operators and vendors are left to innovate on their own by creating new user experiences and revenue-earning added value services.

Dolvane says that one of the aims of the initiative was to avoid the mistakes of the past where technologies such as OCAP and MHP tried to establish standardised ways of delivering a user experience. “It is up to operators to use the code base and stay true to it,” he says.

The companies that have signed up to RDK do so in the belief that any intellectual property they give away will be more than compensated for by their ability to create value higher up the chain towards the user experience itself. “The general feeling is that if RDK is successful that means the industry will be more successful and we will be more successful as a result. A high tide floats all boats,” says Dolvane, who adds that Espial is innovating by creating a fast-performing user experience, including fast channel change and other features that operators are looking for.

The company’s focus on the user experience includes enhancements to performance and aggregating OTT and web content with broadcast video. “It is about bringing in recommendation and the metadata being used for VoD and live TV,” says Dolvane.

“We have been able to take huge strides forward because RDK is transparent and goes down to the system.  In the past you would have to rely on proprietary set-top box software which might not perform well or be well-architected,” he says.

RDK Management’s Heeb says that smaller operators are often to be seen driving innovation. Heeb says that the RDK has to some extent liberated smaller players from the need to follow the larger players in the market. “Smaller operators that have often been held captive by larger operators or vendors are no longer in that position,” he says. Such players can benefit from the use of a common platform by larger players while being set free to experiment at the application layer.

Heeb says that there is a “clean line” separating the elements in TV delivery covered by the RDK, such as tuning, the DRM interface and resource and conflict management, and ‘above the line’ features that could be used to gain revenue, such as the UI, apps, video-on-demand interfaces and so on. While the line is not immovable, the elements covered should not include money-making features, where it should be left to individual companies to innovate as they see fit, says Heeb. “If the operators agreed on a common method on apps and doing purchases they bring it in and it would have to be a common interface but all revenue generating services should be above that,” he says.

For Arris’ Cheevers, the elements that have been included in the RDK are now common currency for operators – the basic building blocks of set-top boxes that formerly had a significant value as proprietary intellectual property but do so no longer. Operators and vendors have plenty of further scope for innovation above that line, deploying technologies that make boxes boot up faster, for example.

Vendors and operator customers still have plenty of options to choose from and plenty of ways to innovate. Cheevers cites the example of solid-state hardware for DVR applications. This is still seen as too expensive to replace mechanical hard drives in consumer premises equipment but could be applied to cloud-DVR services. Operators can also choose whether or not to run a full browser implementation in their set-tops or not.

Cheevers admits that certain elements of the RDK that are currently ‘above the line’ may become standard features in the future to the extent that they will be included in the RDK trunk. “If something becomes ubiquitous there will be pressure to fold it in – that’s part and parcel of things,” he says. However, the revenue-earning elements that differentiate one service from another will always be beyond the remit of the RDK software bundle.

In the case of SeaChange, Hahn says that it can provide additional layers and features on top of the RDK software, including services such as the SeaChange helpdesk. “We also see ourselves as a system integrator. We take different pieces and integrate them with the set-top and take responsibility,” he says. SeaChange is responsible for implementing features including fast channel change and can work with CDN providers to deliver this in the case of IP video providers.

How widespread RDK adoption becomes, particularly in Europe, is still to be seen. Its overall efficacy, versus that of integrated middleware, is unproven. The adoption curve is promising, and a substantial number of vendors are supportive of the overall model. Whether this means faster delivery of new services and a cycle of innovation that matches that of web-based service providers – or whether the complexity of integrating and deploying services in the real world will take its toll – only time will tell.


Read Next