Let’s imagine that you were putting together a Streaming Video Alliance and wanted to ensure you created a better set of technologies to make streaming more efficient and, for operators, less resource stretching. Who would you call first? Think about it.
You have a few options, you either want an operator that has the biggest problem, but also the biggest buying budget, or you want the technology that is most promising or you want a bunch of equipment vendors that understand IP. You may want chip makers that can support better compression on their devices, or encoding companies that can compress it to a smaller stream.
So you go either to the company that will install the most of this technology or the companies who are already doing the best job in technical terms, whether that’s compression, CDNs, routing or backbone and access networks. You may even go to the content companies and ask them to back usage of any new technology, or the largest OTT services or broadcasters that want to switch traffic over the internet. Maybe you’d have to approach all of them to get a “slam dunk” outcome to let your technology win.
So what happened when the Streaming Video Alliance was officially formed last week? We ended up with a motley collection of players with operators represented by Comcast and Charter from US cable, Liberty Global representing cable in Europe, Korea Telecom, Telstra and Telecom Italia from the Telcos and cellcos, Level 3 from US ISPs, Major League Baseball, EPIX and Fox from the content owners and two of the major IP players in Alcatel and Cisco. The other members, like Qwilt, Wowza and Ustream are smaller technology players with a bit of an axe to grind in some part of the network. Yahoo was also there, perhaps because of the strength of its Mobile internet video delivery and Limelight Networks to speak for the CDN contingent perhaps.
A number of articles that came out on the day mentioned companies who were conspicuous by their absence from the list – mostly mentioning one or two names – Faultline thought this week it would go a little further and mention most of them.
First off the absence of the large streaming players, and this right now is Netflix, Amazon, Hulu and YouTube. If those four agreed on a single technology format or protocol then you’d look up wouldn’t you, because it would affect about 70% of all internet traffic at peak times.
The owners of the streaming formats might also be worth a mention – and these are Apple, Microsoft, Adobe and Google – nowhere to be seen, but you could argue that these companies are the problem – in that they sell proprietary streaming systems for adaptive bit rate.
In Apple and also Samsung, you have the operating system controllers of the devices, which would have to be the recipients of the streamed content – both in terms of Samsung TVs and handsets and Apple tablets, handsets and the Apple TV. Not mentioned, but as we said before, perhaps because Apple rules the roost in streaming with its proprietary HLS format. But you’d still call Samsung wouldn’t you?
Another company missing is Ericsson, who can make a huge hole in the mobile streaming market, but which put itself heart and soul behind MPEG DASH a streaming protocol that was supposed to unite all the Adaptive Bit Rate schemes out there. Ericsson also controls almost half the world’s cellular infrastructure, and so might implement this on a grand scale for cellular networks. It has also installed an awful lot of CDNs. Other companies in that line of business include Alcatel, Huawei, to some extent Cisco and Nokia Networks, but only one of those is in this new Alliance.
Surely internet majors would also help and here we are represented by Yahoo! But we are without Google, Baidu, Youku Tudou, Facebook, MSN or Netflix, so there just is not enough internet traffic accounted for here.
EchoStar now happens to own the intellectual property of the company that invented Adaptive Bit Rate video delivery, Move Networks, and it too would find it hard to join a club that is being driven by its arch rival in Comcast. So it’s not there, and that’s despite the fact that it is known to be in the process of creating a new streaming OTT video service, as is DirecTV and Verizon. None of them are invited.
Even Wal-Mart, the other massive shipper of DVDs from Amazon, both of whom have a vested interest to be in on the streaming world. Wal-Mart also has its own online service VuDu.
If this is about streaming to TVs directly, then surely Samsung, Vizio, LG and Sony, not to mention Panasonic or the Chinese TV makers Hisense or TCL. None of them are there.
And as for content, we’re sure that anything to do with streaming, Disney would join, but then again perhaps not if it was being driven by its arch rival NBC, which of course is now part of Comcast. Same logic goes for CBS and Hulu.
The chip community is simply not invited. If you really wanted to help the cellular community handle streaming, you would get Qualcomm or MediaTek involved, ask them to build the final architecture into their chips. Broadcom could do the same for WiFi or set tops.
Most streaming goes through Akamai’s CDN networks, and perhaps it feels it cannot join any club that has its rival Limelight in, because their philosophy in CDN architectures are so very different.
There is no cellular major except Korea Telecom, no Vodafone or Telefonica, China Mobile or T-Mobile or Orange or NTT DoCoMo. And of course no AT&T, Verizon or Sprint.
Finally there are no audio licensing or technical specialists like Dolby or DTS or RoVi or Gracenote, and there are no set top players (except Cisco) such as Arris or Pace or Technicolor, there is no membership for Net top players either, such as Roku.
We think that the membership list tells a little about what this club really is. It is the core of a Comcast strategy for its TV Everywhere architecture. It has no plans to improve the fundamental protocols of ABR (which really do need re-inventing to relieve the cellular community) but instead this looks like a thinking shop for how to get video into the right part of the backbone network for cable companies and perhaps also for other operators. The aim is perhaps to bypass peering and bringing video assets close to local Converged Cable Access Platform servers inside the cable network without the need for more backbone switching capacity. In essence it is a think shop to reduce capex in existing cable networks, so that video doesn’t swamp those networks.
So this nonsense about it being all about an open architecture for online video providers focused on interoperability, sounds highly suspect.
This is about an extension to the CableLabs architecture for IP video delivery in cable networks, with some outside technical partners invited to help. It will not deepen collaboration, because these people already work together with the exception of the spattering of overseas Telcos. It will perhaps redefine the way online video is streamed across fixed networks, but mobile is very much a second thought.
The cellular community has wanted an Quality of Experience method for defining the condition of video when it arrives at a mobile device. There are only two ways of getting at this, from viewing the cellular network end to end, which could be facilitated by Ericsson or Huawei (perhaps Alcatel) or by putting reporting software onto each device (no handset vendors present).
What might happen here is that Cable will adopt a method of measuring this from the network delivery side, which can also work with existing cellular networks like those of Telstra, Telecom Italia and Korea Telecom. But that will not have enough momentum to take in the entire cellular community. We cannot imagine this Alliance coming up with standards for interoperability.
What it may do is present new the IP technologies such as those we have talked about in the past, for instance UDP multicasting, used to wrap and control the bandwidth seeking activities of ADR. It might think about how to place these on cheap cloud servers, which will act as a caching infrastructure to replace CDNs at a fraction of the price.
The rhetoric used to sell this Alliance to the public – that video is about to submerge the mobile world – is not really part of the remit. Instead this is likely to benefit cable operators, making it cheaper to bring IP video to their client base over their existing networks, and it may also be a way of bypassing Network Neutrality directives, taking a different, possibly faster, route for some IP video.
In the detailed objectives it says that it will come up with standards within 12 months and set up working groups in Open Caching and Content Delivery, Testing, Interoperability and Certification, Quality of Experience,
Analytics and Reporting, Security (it has not invited a single security company to be in the loop – but they may well queue to join) and privacy.
If you go back 14 years we can remember that the Internet Streaming Media Alliance was put together by Apple, Cisco Systems, Kasenna (now part of Canada’s Espial), Philips and Sun Microsystems (RIP). It was set up to explore streaming and progressive download and did interoperability and pushed new standards. It was merged into MPEG in 2010 after it had established ISMACrypt and MPEG4 streaming. Chances are this new group will not achieve half so much.