Sponsored By
Once described as the most valuable piece of real estate in the television business, set-top boxes (STB) are now more of a constraint for new service deployment. Most pay-TV providers have a large legacy investment in older, more limited STBs that now restrict their ability to roll out new services and advanced features.
With the advent of virtualization in network functions in recent years, operators took another look at the role and capabilities of the STB and started to explore the potential for the "virtualized set-top box." One of the first operators to do so was Deutsche Telekom AG (NYSE: DT). Managing a portfolio of eight "Natcos" or national telcos for its IPTV service meant that the operator had a large base of 2 million TV subscribers spread across Europe with a broad range of STB profiles to manage. This created a number of challenges for the operator, from slowing new service/feature deployment to creating maintenance complexity.
Deutsche Telekom's Innovation Laboratories were handed the responsibility of developing a solution. Under the stewardship of Dr. Randolph Nikutta, the Telekom Innovation Laboratory has become perhaps the world's foremost source of expertise on this topic in recent years. Telco Transformation caught up with Dr. Nikutta at a recent VTAB meeting, to get his thoughts on virtualizing the STB.
Telco Transformation: What drove Deutsche Telekom to look at virtualizing the set-top box?
Randolph Nikutta: Most operators were looking at the future role of the STB, trying to understand and plan for what we needed. Demands from STBs were rising, we wanted more from the STB than in earlier days. But there is a significant capex impact from the STB. So the question is, how much can I burn on higher-end STBs? It's always a compromise between performance and cost.
Now we need to ask, will customers pay for this box? Can we sell it? What kind of business model can we look at? But these boxes are expensive. In Germany perhaps we can get away with €150 ($168) boxes, but in countries like Hungary and Romania, customers will not pay that much. Maybe we can charge $30-40 for the boxes.
So then how do I guarantee a uniform, high-quality QoE without exploding STB costs? Currently we have 200 STB profiles which are aging, can't support higher-end graphics and are expensive to replace. It would have led to a fragmented user experience with big gaps across our footprint. It would also have affected service development, required more time and resources and couldn't match customers' OTT experience.
And that's not all. The cost of upgrades for clients is also expensive, and very resource-intensive. If anything goes wrong, we have to do a client rollback. So there's more time, effort and expense involved. And with OTT apps now, there are a lot more updates than before.
The STB is crucial, it links a lot of the service elements for us, but we have to look at cost, performance and time-to-market, and find the optimal balance.
TT: What led you to virtualization as a solution?
RN: We spent a lot of time -- a lot of time -- and decided on a cloud browser-based streaming user interface (UI). Part of this was looking at the broader telco shift towards cloud infrastructure. This gave cloud-based UIs a better chance of being deployed, and reap the advantages of the broader shift to the cloud. Everyone was talking about it [virtualization and cloud], so we said, why not the STB? So yes, let's look at that.
Basically we looked at the move to NFV-based infrastructure, and the development of better, low latency networks and thought, it's time to "cloudify" the STB.
But also there were factors such as more demand for advanced UI features, faster innovation cycles with shorter time-to-market requirements. Plus, the technology was maturing; we had a carrier grade, OpenStack-compatible platform.
TT: What were the major challenges you faced in your STB virtualization efforts?
RN: Well, at first there were a lot of myths. People didn't believe we could handle 2 million users at peak, didn't believe support for a local DVR was possible, there would be DRM challenges, etc.
Others took the term cloud too literally! The STB doesn't vanish. We still need hardware to decode the UI stream and the live or on-demand video stream and to decrypt the conditional access protection, but we need less memory, storage, CPU, etc. So the STB can be a lower performing, cheaper box. And we can also use connected TVs.
TT: So what does actually move to the cloud?
RN: So the term "virtual STB" is more correctly cloud browser-based streaming UI technology. This means we move media pre-processing into the cloud. We have an app engine and a streaming engine for preparing the guide interface as a video or picture and delivering it as a rendered output to the STB. And we have media adaptation and a DRM bridge in the cloud.
TT: How has the move to a virtualized STB helped you?
RN: It provides a better QoE to customers. They have a better UI, a better guide. And it is consistent across our deployed footprint, so simplifies product upgrades, new offers, marketing communication as well. It also allows us to move faster -- the time to market is greatly reduced. We can run updates much faster.
Naturally there is also a significant cost advantage. The STB stays, there is no need to replace it, which helps with capex requirements.
This has worked at a number of other providers as well: Charter Communications and Cablevision in the US, CJ Hellovision, SK Telecom and SK Broadband in Korea and Liberty Global in Hungary are all using virtualized STBs.
TT: Why not just go to OTT STBs, such as Chromecast, Roku, etc.?
RN: As an operator, we need control. We are the one held responsible by the customer when things go wrong. Also, for services to third-party devices, what is the business model for us? If we switch to Apple TV and Chromecast, we don't own the customer data, we don't own the start screen.
For multiscreen, we accept we can't avoid it. But with the STB, we can control it. Also, we are working with connected TVs. Now Samsung, LG, for example, are offering us operator apps on the connected TV. We can have an operator portal on the TV, which is more interesting for us.
TT: What areas of development are you now looking at?
RN: There are still some questions to be answered. We are looking at how we might be able to develop a cloud UI for hybrid TV services [where services are delivered over multiple different networks]. And we are also interested in TV services in mobile-only countries, how that might be managed.
— Aditya Kishore, Practice Leader, Video Transformation, Telco Transformation
