Cloud migration was one of the standout themes of 2020, with many operators and content providers rushing to get remote capabilities online. Now that the dust has settled, more nuanced approaches to cloud migration are becoming the norm, and a panel at this year’s digitized Anga Com aimed to explore this more measured approach.
First up, we heard from Synamedia’s CTO of Video Network, Nick Fielibert, who argued that cloud migration makes sense for most cable operators, just that the extent to which this is true is highly case-dependent. As most of our readers will know by now, some things make sense purely in the cloud, some on-premise, and many use cases benefit from switching regularly back and forth between the two.
Fielibert noted that video processing technology is advancing at ever greater speeds to keep up with an explosion of content, which requires proper management via software, containers and microservices architectures. Trends such as the non-linearization of video consumption and growing consumer demands for broadcast-level latency as this happens are only adding to the pressure.
Amid this turmoil, cloudification can offer agility in workflows for all types of content, global reach, and more flexible business models as many operators move from a capex model of costs to an opex one. Fielibert reminded everyone that flexibility is the key argument for cloud migration, as the time to market for new workflows could be at “the snap of a finger”, depending on the system, and there is potential for near-infinite scaling – reducing the need for upfront capacity planning.
However, he cautions that cloud servers tend to be more costly than their on-premise counterparts, largely because they are a managed proposition which will almost always be able to back-up any failures. Most of the costs comes from data egress (ingress is usually free), as most cloud services charge per GB, not per GB per second. All this could pose difficulties for operators that want to maintain their typical broadcast-type service level agreements.
Fielibert told the panel that 24/7 broadcast use cases rarely benefit from an entirely cloud-based environment. “The only advantage is global reach. If you are a cable operator distributing only to subs in your own network, it makes less sense,” he said. However, he notes that the cloud still does a stellar job as a back-up in case the on-premise servers go down, carrying the torch until the missing head-end is fixed.
Another huge draw is being able to instantly test out new capabilities and technologies such as codecs, scrambling methods and applications. If a certain capability seems like it should run 24/7, operators can easily move it to an on-premise server.
Next, we heard from live workflow provider Appear, whose Chief Information Officer, Per-Henning Almvang, took us on a whistle-stop tour of the encoding and transcoding jungle. Presenting a “simplified” flowchart, Almvang illustrated how the full picture is far more mangled.
For instance, at the transcode/encode stage, a hybrid operator must decide whether to do a single round of compression for IPTV and OTT, or whether to adapt content requirements to tackle them separately. Hybrid compression – the former option – reduces source management and decreases coder capacity requirements, but the trade off is that it reduces bitrate efficiency.
Not only does this increase bandwidth costs, but reduced bitrate efficiency potentially creates a bottleneck, as the entire network must run at the bitrate of the weakest link in the chain. High performing codecs may be out of the question.
“Ultimately, operators need to provide maximum quality at best cost,” Almvang told us. But how to best keep costs down? He argued a holistic view of the workflow is the answer, as only then could operators make well-informed decisions about restructuring their total cost of ownership (TCO).
“You need to devise your goals and be open to multiple systems,” he continued. “There are multiple challenges in the value chain simultaneously, but these could be easily solved by one simple system.”
A helpful tool for deciphering priorities is the so-called trade-off triangle shown below. On the three corners of the diagram are the three factors – quality, latency, and density – which, if focused on, will usually be at the detriment of the other two. Almvang argued that latency would likely be the priority for sports events, but other entertainment may prioritize quality.
Almvang notes that more wiggle room inside the triangle comes from choosing a more efficient codec, but this likely comes with upgrade costs to both the delivery network and end-user devices. Therefore, an operator must decide on their priorities to ascertain whether this wiggle room is worth loosening its purse strings.
Almvang echoed Fielibert’s prescription for cloud-migration – great for events, global reach, and dynamic scaling; less good for keeping down TCO, especially for 24/7 use cases.
Finally, we heard from IPTV vendor Zattoo’s Head of Sales, Julian Hems, who tried to set straight many of the anxieties that operators have aired behind closed doors when it comes to migrating pay TV services to OTT apps. The well-meaning talk seemed to expose more wounds than it healed, although we can certainly see how this is still a worthwhile conversation.
Hems told us that many operators fear their backend and middleware is not up to the job of onboarding multiple app experiences. He says the problem is largely rooted in the siloed worlds of IPTV and OTT development, with vendors in both camps often not willing to talk to one another.
“If you want to go app-based, you need to break up these silos,” he warned, arguing that the customer ultimately does not care about delivery technologies, and just wants to enjoy consistent quality and functionality across all devices.
Another glaring obstacle is antiquated OSS/BSS systems, many of which are not flexible enough to book additional devices and apps. Hems argued that legacy OSS/BSS is the main bottleneck squeezing operators’ time to market for new products.
It seems that while every vendor and analyst is preaching the promise of OTT, operators are still worried that the ‘content reality’ might not yet meet their needs. Hems says this is not unreasonable, with all kinds of creative restrictions on rights to recorded content making it hard to manage titles across multiple applications and player technologies.
This is further complicated by regulations for ad integration via dynamic ad insertion (DAI) and substitution. If operators want to be at the helm of these promising revenue streams, they must juggle various devices and applications here, too.
Hems also confirmed that operators have lost much of their swagger in the face of OTT giants and big tech players like Google dominating much of their content and UI. “Many ask us, ‘as an operator, am I big enough to talk regularly and make good deals with all these players? Will they even talk to me?’” Helms told the panel.
Although he certainly brought more problems than solutions, Hems argued that once these various challenges are tackled, “appification” of a pay TV service is a promising opportunity for operators. Not only is the reduced reliance on hardware one less logistic to worry about, but there are new business opportunities through OEM partnerships, while brand engagement is increased.
“There is more customer engagement where it matters. Instead of having your brand hidden on a set top, it is now on a home screen of every customer device – a stickiness that is much more important nowadays with the huge choice of streaming apps,” Hems argued.
On the reduced reliance on hardware, Hems argues this would also ramp-up multiscreen capabilities. While organizing second and third set tops in one home is both costly and timely, an app allows you to instantly access every screen in a home.