Killing ‘OTT’: Why the Term 'Over-the-Top' Has Lost its Meaning
For almost 20 years, I’ve not only worked in the streaming media industry, but I’ve also written about it continuously. I might spend more time writing about the sector than any other single activity—voicing strong, sometimes testy opinions on architectural strategy, educating neophytes as to the way the industry failed in its previous three attempts to do what they think they have just “discovered,” or even contracting technical phrases to coin new words (such as “CellMux” for cellular multiplexing in 2010 and “TransMux” for transcoding-multiplexing in 2007).
So the words matter to me. I get all antsy when a term is being misused. For several years now, when I’ve presented on virtualization, I’ve always ensured that my audiences were clear that the widely misused term “cloud” is an economic term, and not a technical one.
Outside of streaming media, I also find my hobby, “drone flying,” has suffered terribly. We fliers used to call our hobby “quadcopter flying,” but the media decided that “drone” was far more likely to sell fear, so now my model-flying hobby is being run to ground by the thirst of the press for a “Drone Strike Takes Down Passenger Jet” headline. It is like it has been misappropriated for the media’s benefit, and it is damaging the hobby.
But my current rage is going to be focused on a term that is widely, horribly, and terribly misused at the moment. That term is one that you probably used this week. It is a term, like “drone,” that has snuck into the vernacular so deeply that even I have found myself using it recently to generalize about streaming technologies.
But if I turn my pedant mode on, the term (I can barely even type it at the moment) “OTT”— which stands for “over the top”—is undergoing a notable semantic change. That is, in its own right, ultimately not a problem. I have no problem with reusing a label, so long as it is not used to mean both its original meaning and its post-change meaning at the same time. However the blur caused during this particular change is causing people to talk at cross-purposes without realizing it. And that is a problem, particularly in a technical services industry like online media.
Let me explain: OTT is being used in our sector to encompass anything today that is not “traditional broadcast.” It is being positioned as the alternative to some other form of broadcasting. Two apparently opposing engineering strategies are being pitted against each other, due in part to the media’s desperate hunt for a newsworthy fight.
In actuality, OTT itself refers to the “placement of a paywall outside of the subscriber network.” As I advise operators on their architectures, I always predicate any “OTT Service Architecture” session by clarifying this point.
When the term started to appear, it emerged against the “walled garden” model that was prevalent in then-emerging voice-over-IP (VoIP) services. Essentially, what happened in the traditional voice networks was that some organizations were paying a local telco for a subscriber line to be physically installed in their building, and provisioned with IP. They would then contract a third-party provider of IP telephony services to terminate their calls into other telecoms networks, and they would pay the third party for those VoIP services directly, with the telco making no revenue for the service. This meant that the VoIP payment was going “over the top” of the otherwise “walled garden” of the local telephone network. By aggregating the voice traffic and buying wholesale interconnects into other telephony networks, the VoIP providers made it cheaper for the customer to make phone calls, which was seen as erosive to the local telcos’ core voice telephony monopoly.
That was what “over the top” was born meaning. And in purist circles of experienced engineers, that is still what it means. For example, Skype is a classic OTT service.
Focusing on video, Netflix is always put on a pedestal as the archetypal OTT service provider.
Netflix does indeed collect subscriptions— generally—directly from consumers. So it is indeed an OTT model. But most usage of the term “OTT” with regard to Netflix focuses on the model that it delivers the video over the top to the subscribers, and this is not actually always the case. While Netflix may, in some cases, deliver in this way, it does not, in practice, always deliver video to the consumers over the top into the subscriber network. Increasingly, Netflix provides its Open Connect technology to ISP operators, which internally host servers specifically for Netflix within the subscriber network, where the operators can control the quality of service between the Open Connect system and the end user on their own managed network.
In terms of delivery, this is architecturally exactly the same as IPTV. In the IPTV model, the subscriber adds video services to the subscriber package, and these video services are delivered on-net from servers within their own network, allowing the operator to guarantee services. This is the same for fibre and copper telecoms’ IP networks as it is for HFC/DOCSIS “cable” networks with ISP services. Both the Open Connect and the IPTV architecture receive a contribution feed from outside of the network, but the delivery is done on net in both cases. This means that the subscriber network can absolutely manage quality over its managed network links, and this important differentiator is lost by sweepingly calling Netflix an OTT service as if it were the same as, say, the BBC iPlayer services or even YouTube’s services (which typically do not cache their content within the subscriber networks).
So the only technical nuance between Netflix and the subscriber ISP/telco’s own IPTV service is that Netflix directly handles the user subscription from outside of the telco’s “walled garden.” And that line gets blurred when the cable provider bundles Netflix with its subscription: at that point Netflix is, for all intents and purposes, no longer an OTT service in any way, but is in fact a pure IPTV service.
So any conversation about OTT architecture that is about anything other than paywall location is immediately ambiguous.
To highlight this further, where OTT has increasingly been used to represent “IP-delivered video service” as distinct from IPTV or broadcast TV, it is also key to appreciate that much of the backend workflow producing “broadcast video” is also IP-delivered already. In fact, there is increasingly little left in any video workflow that is not being shipped around on IP.
Why am I so heated about this?
I believe it is key that we do not think of “broadcast” and “IP video” as different things any longer. If we do, we will architect in isolation and this will hinder our ability to keep up with our customers demands. The two will benefit best if they grow together, and not apart.
The consumer is moving to IP. Whatever we call the delivery model, it has to be good enough for the consumer to consume. If we try to create separation by bundling too many things into terms like “OTT” and “cloud” (etc.), we will only make that isolation between the schools of thought more divisive and less constructive in the architectural planning that goes into meeting consumer demands.
So please use your terms carefully: if you want to talk generically about IP-delivered video, then perhaps choose “streaming video.” If you want to talk about where your subscriber billing is done, then talk about OTT. And if you want to talk about quality delivery on-net, then talk about IPTV. Where the lines blur— as they do with Netflix—perhaps talk about “hybrid IPTV/OTT,” but do not simply assume that it is OTT. Or perhaps it might be nearly time to drop all the distinctions and simply call all of it “broadcasting.”
[This article appears in the Summer 2017 issue of Streaming Media Magazine European Edition as Killing "OTT"]
OTT is growing worldwide, but trends in viewing habits, monetization, and media workflows vary from region to region.
New survey asks industry members to weigh in on the state of OTT today and share their insights into where they see the market headed in the future
The broadband-based service will be rolled out first in the UK, then in Italy, Germany, and Austria. There's no word yet on whether it would offer 4K content or include bandwidth to the home for 4K streaming.