Publié par Laisser un commentaire

(English) Live OTT streaming – Industry feedback from CDN World Summit London 2014

Désolé, cet article est seulement disponible en Anglais Américain. Pour le confort de l’utilisateur, le contenu est affiché ci-dessous dans une autre langue. Vous pouvez cliquer le lien pour changer de langue active.

Last week I led a round table on the future of live OTT TV and it’s implication for CDNs during the last session of Informa’s CDN World Summit in London.

My first point to open the debate was on QoE. I pointed out that mobile telephones are a giant step backwards in term of voice QoE and service availability compared to good old fixed lines. However we’re all happy to renew dropped calls lose coverage or ask our correspondents to repeat on our mobile phones because we gained so much more than service QoE with mobility. I then suggested users might accept a similar trade-off and embrace lower QoE for OTT TV than broadcast, in exchange for lower costs, mobility, greater choice and personalisation. The reactions around the table made me think I’d just insulted the queen. There was emphatic disagreement. TV is TV and will always be TV said the TV operators and nobody dared take issue. I guess that’s what happened in the boardrooms of railway companies when airplanes arrived. One of the non-TV-operator participants did agree that maybe - except for sports - QoE might be traded-off for greater choice. At this point, the challenge of content navigation was brought up for search and recommendation.

That got us talking about “long-tail live TV” and if it might ever makes sense, i.e. being able to watch a unique live stream that you really care about. That access might make you so grateful that even if the quality wasn’t always pristine you’d still be happy. This idea is buoyed up in an OTT rather than broadcast context. Indeed all the TV markets I’ve worked in, even if they have many hundreds of channels available, invariably have 10 or fewer channels that any one community is prepared to pay for. One of the key promises of OTT is to abolish markets, typically under a satellite footprint. All those start-ups targeting Diasporas are going to find tough competition as the big guys come into their nascent markets more and more.

From a financial modelling point of view, the satellite broadcasters around the table were pretty excited about the fact that for live OTT, if you have a tail-end channel that nobody is watching, your Opex goes down to zero. This for them was the real opportunity in live OTT.

Consensus was easy to get on the fact that live OTT TV brings mobility, however nobody was clear yet about a killer use case where this is really important. Watching videos on the tube or train is still very much a download experience and rarely legal at that.

When I brought up the question of when rather than if, Netflix starts live streaming nobody felt ready to pick up the gauntlet. I’ll keep that for another day.

Our debate wound up over an interesting discussion on the blurring of boundaries between linear and on-demand content. Typically a shopping channel can be played out from an automated server with people being able to interact and turn a multicast stream into a unicast one. The final feedback from two operators round the table was that Multicast is only really a panacea for large Telcos that own a network. For the rest of us the cost benefit analysis turns out much worse in the real world than on the drawing board of business planning.

This left me with the clear impression that there are still problems out there looking for solutions, not the other way round for a change. As many network and service operators want to build their own solutions rather than relay on the global CDN operators, we'll probably see a major player emerge from the likes of Anevia with its edge caching, Broadpeak with its Nano-CDN, Media Melon with its QoE analysis or Octoshape.

Publié par Laisser un commentaire

IP&TV World Forum MENA report Part 2/2

Désolé, cet article est seulement disponible en Anglais Américain. Pour le confort de l’utilisateur, le contenu est affiché ci-dessous dans une autre langue. Vous pouvez cliquer le lien pour changer de langue active.

This second report form this year’s IP&TV World Forum in Dubai covers in detail the Selevision solution for OTT and operator branded IPTV services. There is also a description of Mariner Partners xVu product demonstration and Anevia’s adaptive bit rate solution.

Selevision
I visited Selevison’s booth, the largest on the floor except of course for Etisalat that had its own section. Selevision is a Saudi-based company with a complete TV solution for operators and also a B2C offering aimed at 15 Arab speaking countries. The core TV product has all the latest bells and whistles for an STB based solution but I didn’t see any multiscreen on the booth.

Bilal Abo Alul, the Bahrain project manager for Selevision, was my guide.
The main product on display that he showed me was a hybrid DVB-S / IP HD box with 2 tuners. The device is manufactured by Strong and has the Oregon media browser on board. Push and pull VoD are available and the PVR supports time-shifting with its 250GB HD.

An unusual feature is that the circular time buffer stored on the disk is 3 hours long instead of the usual 1 hour and more importantly, it isn’t reset each time the channel is changed. I’ve often cursed my live-TV rewind because it couldn’t do just that. I suppose the ultimate solution would then be - to somehow link this with a record and a catch-up feature. You could then zap all over the place looking for the best movie, once found, hit the start-over feature or store it in your library. I know some platforms do bits of this but it’ll be quite a while until the seamless user-experience I just described can be delivered, if ever content-owners allow it. Selevision pointed part of the way at least.

The box gives access to a content catalogue aggregated by a third party, Grey Juice Lab. The VoD store already has over 200 titles with 500 being the short-term target. The Technology is assembled in house with components developed in France by Vianeos and Hyperpanel.

Selevision has already sold 20k boxes in Saudi Arabia. The box comes by default with VoD and Free To Air channels, and users can also subscribe to Al Jazeera sports and Abu Dhabi media packages. Bilal told me that OSN the leading TV platform in the region should be coming soon to the box.

On the IP side, Catch-up services over 7 days are available for 10 channels. Currently this is on an all-or-nothing basis per channel which could prove problematic in the future as some stations agree for only parts of their programming to be made available in this way.

Other interesting features I was shown included RSS feeds and the unavoidable YouTube App. Note that for the latter, no censorship is required because the IP feed is local and therefore supposed to be censored itself. Surprisingly content can be put on an external hard disk. Basic media-centre features are also available to consume content from a USB stick.

Bilal also showed me the STB and service that they put together for the Bahrain telco Batelco based on an IP-only version of the same middleware. 160 live TV channels are available (3 FTA and the rest within different Pay-TV packages). Also on the Selevision booth was a customized version of the box for KAUST (University) on a Motorola platform. 10k users are deployed currently on this slightly older version.

Dubai Studio City

Dubai Studio City had a booth that was promoting the free Zone providing office space and other logistics and legal aid for companies wanting to come and do business in Dubai. In the same booth, Rufoof (which means ‘shelves’ in Arabic) was also showing its wares. They are a 10-person start-up claiming to be the first local platform to assist publishers on electronic distribution on the latest devices. They do the software development and Content Management system and if I understood correctly, they even act as an Electronic publishing house themselves.

Anevia

This French company that was here to talk about … [drum roll … wait for it] OK so you guessed: OTT & Multiscreen. I say talk about because they didn’t bring a demo, it was just PowerPoint. However their solution seemed interesting and you can tell that they have been thinking hard about the whole adaptive bit-rate issues for a long time. I’m actually a bit miffed that their CTO brought out a white paper on the subject just before the one I authored for Verimatrix and Harmonic came out last year.

[Warning technobabble coming ⇒] The architecture they presented was more resilient and scalable that most of what I’ve seen so far. The ABR multiscreen solution uses a circular buffer just before the chunking takes places and the Origin server gets its feed. That way Anevia can easily implement a scalable catch-up TV and NPVR or Time-shift feature in the network. The Edge servers in their complete solution are offered by default with 10 GB of storage so that up to 10k subs (assuming 1Mb streams) can be supported at each edge server. They also recommend putting their Balancer at the edge for load-balancing and edge-based failover.

The Anevia in-house monitoring solution uses data collected from the Origin server, the edge servers and the load-balancers, and also from optional probes that can emulate user behaviour to capture real Quality of Experience.
[⇐ All clear - end of technobabble warning]

Although so far Anevia has sold all the separate components of the complete OTT, multiscreen solution in various combinations, only one instance of the complete solution is being deployed. This showcase operator deployment will be announced soon.

Mariner Partners

This Canadian company was showing off the latest version of its xVu product. Mariner is a company born out of an operator with founders that understood that the traditional Telco QoS (Quality of Service) approach to TV would not cut it. QoE (Quality of Experience) is the mantra of many of the suppliers in this space, but Mariner brings an interesting touch. They approach the issue from the Customer Experience angle.

All the QoE vendors have something to offer to all Telco stakeholders, but Mariner’s approach is different in that their xVu product is also designed from the outset for the customer support teams as well as for the Network Operations Centre (NOC). The impetus here is first on preventing and fixing a bad customer experience more than running a smooth network. A typical feature that illustrates this approach is the ability to make sure the technician doesn't leave the customer premise before he receives the green light, so if there must be a truck-roll at least it has the best chance of fixing the problem.

The main constraint on the Mariner solution is that all the devices in the network must either use Microsoft Media room, Cisco’s VQE (Video Quality Experience), Minerva Middleware or be compliant with the latest TR-135 protocol. If none of these criteria are matched, a software agent must be embedded in the STB. I theory this is relatively trivial, but real world project have a tendency to veer away from theory. The demo shown in Dubai was impressive and seemed to make this constraint worth the effort. I was shown how operators can drill down from a global network view through all the layers of the topology to a specific customer STB that has an issue. One unique feature of this monitoring approach was the ability to identify on the silent errors: in IPTV architectures that include packet repair, customer quality may seem satisfactory, while the packet repair is actually burning the CPU faster and faster. The operator is able to fix such silent issues before customers complain.

Dune HD
Sometimes I just don’t get it. What did Dune HD hope to get out of the show? On their booth, this German manufacturer was showing Hybrid STBs and media players. They come from the retail sector and I think I understood that their products’ USP is that they can also be used stand-alone. Hmm, maybe they should then present at a CE show. Anyway their key customer for managed services is Kartinal TV in Germany.

Dune HD claims to develop only hardware and software. Their devices “don’t need” middleware. The low-level (i.e. non-HTML) aspect of the UI means it’s pretty nifty even on low-end chipset. As usual with this kind of setup, only one chipset manufacturer is supported; in this case it’s Sigma designs. The boxes run under Linux and are made in Taiwan.