Your browser is not supported. Please update it.

18 September 2014

“Set top boxes disappear, right?” The age of the cloud hits IBC

IBC very much had its focus on the cloud rather than 4K this week, at Europe’s largest digital media conference and exhibition. And Paul Larbey of Alcatel Lucent’s Video Business Unit put it best when he said, “Set tops boxes disappear, right?” as his starter for ten, as he began a far reaching presentation that quite excited Faultline – at first.

As we told Larbey it is our job at Faultline to say outlandish things like set tops now die, and to hear it from an established IP vendor was very much a turning of the tables. But other announcements at the show also underlined the emergence of cloud strategies, including within areas which we do not usually cover, such as workflow and playout, something for instance addressed by Imagine Communications, once the broadcasting arm of Harris, which now consists entirely of cloud software and services.

But there are more cloud clues in between many of the comments that we heard at the show, in particular the quiet demise a few months back of the XG5 specification for Comcast. At a casual chat with Charles Cheevers, the Arris CTO for CPE Solutions, we were told, “You must stop referring to the Comcsat XG5, that is no longer an active program,” a move we could hardly believe.

One of our constant concerns last year was just why was everyone going to bigger and bigger CPE devices in 2013 as we approached the age of the cloud. Certainly AT&T, Comcast and DirecTV had all specified giant multi-room DVRs with onboard transcoding. It turns out that this was the last hurrah for large in-home multi-room DVRs, except perhaps for DTH players. Comcast we understand will now continue to roll out Xfinity X1 along with versions of its improved UI and WiFi (dubbed X2),  but it will never reach the headless media gateway that was specified in January 2013 as the next big thing. Both Pace and Arris had developed reference specs for it, which they showed as recently as this January at CES, but now the cloud is the new direction and both will have to think again.

What we think they will do now is more or less nothing for about 3 years, and wait until X1 technology is commonplace, and then Comcast will scale back its CPE. By then its network will be largely IP, and it will run both a massive CDN and a contiguous WiFi cloud (these days they are calling it a canopy) that comes from the 21 million broadband homes, having their internal WiFi used as a commercial hotspot with a second SSID, or should we call it a homespot.

From here the technology is a cluster of IP standards, and if the interest that Comcast, Time Warner Cable and CableLabs has shown in a recent DLNA guideline, CVP-2, is anything to go by, that will more or less define its future CPE, potentially with a retail product that consumers can go out and buy, rather than one funded by the operators.

CVP-2 emerged at the IBC event with a new name, Vidipath, preferred as a deployment feature inside existing gateways, but also an entire product specification in the future. We wrote in March that CVP-2 or Vidipath will extend pay TV content sharing in the home.

It will natively support DTCP-IP security and an HTML5 UI. And extend to digital TVs, mobile devices, Blu-ray players and games consoles. Video will be in the MPEG DASH format and if Apple ever wants to play here it must embrace MPEG DASH ahead of HLS. This is the first time that the DNLA has concerned itself with subscription TV. It will have remote diagnostics for CE devices, to reduce help desk calls due to complex hardware or services. UPnP Energy Management and a Networked Device Power Saver feature are other parts of the CVP-2 guidelines that will ensure devices conform to new energy regulations regarding sleep-state power consumption.

Certification for the specs under the Vidipath name are now open and first certifications should be out late in the year. It will take a while before a device comes along with the full DLNA stack that is ready to work as a cable delivery service, but this all implies that the major MSOs are headed down the cloud route for content, with a full IP path. And the DLNA tells us that already 15 member companies are preparing for certification and products are expected on retail shelves Q4 2014. Our take is that these will not be home gateways that can be plugged into the cables of major MSOs complete with Conditional Access, but that will one day come. Overnight that might mean that the humble set top is not only dead, but it is replaced by a retail product not funded by cablecos.

All that would be needed would be a service rather like those being prepared by Alcatel, Cisco or Ericsson, to handle Adaptive Bit rate delivery and conditional access, run from the cloud.

So let’s go back to Alcatel and Paul Larbey. Larbey works within the IP Routing & Transport Business, and mostly wants to talk about the cloud, or better still CDN, but began by striking out and writing off the one product his rival Cisco makes, but which his company is not currently engaged in – the humble set top.

Three years ago, a series of mad claims led to US national newspaper journalists claiming all set top sales would stall, and they called us and asked why the set top had any future, and Faultline spent a few hours explaining that while the CPE was almost always 50% of the cost of any given network, that there were many new functions like TV Apps, or HEVC decoding as well as transcoding, that were crying out for admission into TV set tops, and that this processing had to be done somewhere, and while there was a box already out there, at a given price, it made sense to add to its power.

Had some smart Alec come back to us and said “can’t all of that be done in the cloud” we would have said that once the cloud has scaled and there is a widespread acceptance of distributed network video appliances, that this was virtually the same as the set top, likely to cost nearly as much. That was then and this is now.

Larbey began counting things on his fingers – 1) “There is a shift to IP video” – we agreed, detailing some timescales for cable operators as they deploy CCAP. 2) “The network needs to be more flexible,” we just shrugged, that’s always true. 3) “There are never enough people in a cellular cell to warrant the MNOs investing in multicast,” by which Larbey does not mean that LTE Broadcast doesn’t happen, it’s just that it’s only for Live events. 4) “There is a shift to personal devices,” and here we nod enthusiastically, this we too have predicted, and these do not plug into a multicast.

5) “There are more personalized TV line-ups, which means more Unicasting. And it can be quite complicated for consumers to find content that they want which is on right now.”

At Faultline we don’t like to agree with ANY vendor and in particular never a large one that claims to know it all, so we’re sitting there about to go through a list of our own pro-set top points, when Larbey hits us with the big one. “We have done a number of studies for a number of operators based on the viewing figures that they have today. In each case they wanted us to work out what it would cost to replenish their networks to support their pay TV environment, and in every case going cloud was by far the cheapest.”

“An all unicast system is simply cheaper today,” he finished, and this has been true for at least the last 2 years, whether you are an IPTV player a DTT player or a cable operator. For a DTH players their cost is already sunk, so that’s different. For DTT about 25% of the most watched content is still worth broadcasting, the other 75% needs to come from the cloud.”

It slowly sinks in that what Larbey is saying is that he hasn’t really done a study, but he has gone a lot further, and answered an RFP (Request for Price) and has bid for and in many cases won, an all IP shift from where operators are today and he won them based on a unicast cloud based architecture. After all he sells an elastic CDN with accelerated video, with Virtualized servers and multiple layered storage – solid state for poplar VoD titles and disk too.

The Alcatel Velocix platform already holds 28 petabytes of data, or 1,000 years of HD video which is the equivalent of 50,000 TV channels in any given week, or 9 times the Netflix catalog.

Alcatel’s strategy is simple – simplify the home, because that’s the hardest place to get an engineer to fix anything; optimize and virtualize the network and migrate to the cloud, where it already has cloud DVR. It’s not unique in offering this type of lineup, but it IS unique in having convinced 8 major tier 1 customers to go down this route across 12 countries and right now it can only name 3 of them in Telecable, Telefonica and Swisscom out of those 8.

Now Telecable hardly counts, it is a small Spanish cable operators with about 150,000 customer homes, but the others are clearly first tier. We challenge Larbey on the ownership of Telefonica as a customer, noting that Ericsson is supposed to manage its Global Video Platform and very carefully Larbey says “We are the incumbent supplier in Spain, Telefonica may have deployed other technologies elsewhere, but not yet in Spain.”

It is here that he lays claim to the turnaround at Telefonica on the Imagenio TV service. This has languished for years having got to just under 800,000 customers in 2012, when the Euro zone crisis hit, and all the pay TV suppliers in that country fell on hard times. It fell to almost 600,000 users. “Today it is in 1.2 million homes having jumped 91% in the last quarter,” said Larbey. Alcatel has helped it add nPVR and a new UI, and invoked faster channel hopping. Other features that have been added include recording an unlimited number of channels simultaneously, into up to 350 hours of network DVR storage, as well as full catch up TV, working across 130 channels, automatically running for 7 days. There is also a mosaic TV feature where viewers can select between six different views of a live sporting action with instant view change.

Swisscom too has won a few plaudits for its TV 2.0 which took it out of Mediaroom, and perhaps the long awaited shift to Mediaroom inside Telefonica will indeed only be confined to the Brazilian and other LATAM operations, now that Telefonica has seen this turnaround.

If there are 5 other customers yet to be named, all of which have these facilities, then Alcatel is about to become, once again, one of the pay TV leaders, this time instead of that being on the back of Microsoft’s Mediaroom, it will be on the back of its own efforts.

We are fairly sure that Comcast is one of the other unnamed customers because we spent part of last week hearing from Alcatel how it was also managing carrier class WiFi, and that this involved a customer who had put on 2 million hotspots in the US this year, which can only be Comcast. By the time it has named all 8 of these customers, the pay TV market may well be a very different, very cloud oriented place.

Even Ericsson tried to get in on the cloud act at IBC, but as usual it is a touch behind everyone else on any subject that is not cellular.

Ericsson announced at IBC what appears to be more or less the same software that Alcatel has already installed at Telefonica and others. It calls this MediaFirst and positions it as a kind of TV Anywhere system for Mediaroom, essentially CDN delivered multiscreen, with Cloud DVR, set for launch in the middle of 2015. It has added personalization and a big VoD catalog and says that this is already in trials, there is real code and hints that Q2 will be the time to name first customers. Well it had better deliver Telefonica and Deutsche Telekom among them, or there will be little left of its Mediaroom acquisition worth having.

It said that delivery would be just as good on public and private clouds supported by a unified backend that empowers rapid release cycles and a common UI throughout the platform, no matter what screen it comes on.

We’re pretty sure that somewhere in our notes taken over 6 days at the show we heard who provided the recommendations but cannot lay our hands on it, but iWedia was named as the UI provider.

It also underscored this with the acquisition announced at the show for Ericsson to buy Israeli network DVR specialist Fabrix Systems, a cloud storage player for video applications, which will surely power its cloud DVR. Ericsson paid $95 million for the company that has 103 employees and the deal will close in the 4th quarter.

While the cloud clearly has winners and Alcatel may be one of these, Cisco and Ericsson may not be too far behind, Cisco is clearly having to think once again about what happens to its own set top business if US cable decides that the set top has had its day. Certainly the old Scientific Atlanta business is part and parcel of shifting the US cable operators to CCAP and all digital networks, as is Arris. But companies like Pace have little to fall back on except the home gateway, if the set top shifts into the cloud, and we would need to see from Pace and a whole list of other companies, a credible post-cloud strategy, before we could cheerfully see its future beyond the next 2 to 3 years.