Best of Web: OTT Trajectories and Cloud Computing
Mark M. Myslinski | OTT Executive Magazine | Fall – 2017 Issue, Pg. 18
Companies competing in industries based on video compression and processing are at a tenuous point today. This is a time where those companies will have to make a big fat choice as to whether to invest in appliance based capex that is frankly getting dirt cheap, or in a cloud-based model that is opex-based. These are two very distinctly different types of investments and difficult to compare in total cost of ownership (TCO) models. Yet as a company competing in this arena over the next 3 years, you will find yourself sitting in one camp and competing against companies in the other camp. And these technologies are so distinctly different that someone will likely have a distinct advantage.
As you contemplate moving from depreciation-based capex models, to overly opex-based models and considering TCO, what I would say is that identifying some of the trajectories taking place within these technologies could provide some clarity for your path forward. I say this specifically within my current strong suit, over-the-top (OTT) and broadcast television, of which I continue to cultivate the joining of the two. And here I will say that today we are not completely there yet, but this discussion gets us a lot closer.
Cloud Resources Today
Our discussion of cloud infrastructure needs to address two components, the application software and the underlying compute resources. And today, the cloud cost curves for video processing in general are heading sharply in a good direction, but it is elements of the underlying computing resources that need to be addressed for the TCO model to be truly effective.
Cloud application software is referred to as Software-as-a-Service (SaaS). The mainstream SaaS application software that covers storage, off-line encoding, Adaptive Bit Rate (ABR) encoding, cloud DVR; all are coming into line to make the TCO model of cloud-based solutions effective versus that of traditional appliances.
It is cloud-based Constant Bit Rate (CBR) linear encoding in particular that lags effectiveness in the TCO model. Here, most of this is due to a) the costs of computing resources, and b) egress costs. These computing resource costs become reduced as the volume of encoding resources is increased. The primary way you might increase this volume would be through the addition of local systems to your plant (here plant can be cloud virtual), or through the increase in consumers such as by offering off-net OTT services to public consumers in addition to your cable subscribers.
As for egress costs, these are the costs of sending media from the cloud to, in this case its local headend destinations. These costs can be quite substantial, although some commercial cloud providers are starting to cap egress costs based on volume. And if you’re wondering, the act of sending media to the cloud is called ingress, of which many commercial cloud providers today offer cost-free ingress. Hmmm…
As for the cost of the SaaS application software for CBR linear encoding, these costs today are manageable in a TCO model, and will be getting even better as vendors are starting to enable their SaaS to run on multiple commercial clouds (called multicloud). This leads to price competition that will make the TCO of the overall cloud-based model even more competitive.