ABSTRACT
Recent years have seen the emergence of Serverless computing, a model in which engineers deploy software into the cloud, but leave it to the cloud provider to provision and manage underlying servers. This enables delivery teams to focus on achieving business objectives with fewer distractions.
Through use of a fine-grain Pay-As-You-Go (PAYG) model, it also enables better matching of capacity costs to real demand.
In this paper, we explain why the Serverless model is particularly attractive for TV services, and how it can reduce operational overheads and lower barriers to entry. We start by covering the key concepts underpinning Serverless, with specific reference to compute, database and media encoding services.
We then show why TV is well-placed to benefit, basing this on the demand patterns that TV places on IT resources and drawing on the concept of elasticity of capacity. For illustrative purposes the paper features a specific Pay-Per-View (PPV) use-case.
INTRODUCTION
Cloud computing has grown rapidly over the past decade, with public cloud providers such as AWS, Azure, and GoogleCloud firmly establishing themselves as viable infrastructure platforms in multiple industries.
Numerous TV service vendors and operators e.g. Izrailevsky, have embraced cloud with great success, enabling them to deliver highly scalable and available solutions with global reach, and to manage costs by paying only for resources they consume on a PAYG basis.
With the concept of Infrastructure-as-a-Service (IaaS) firmly established in the industry, cloud providers are increasingly promoting a Serverless approach which, at its simplest level, means that cloud customers do not need to be aware of the servers that underpin their services.
In this paper we explain why Serverless is an attractive paradigm for many TV use cases, and how it has the potential to transform the industry by lowering barriers to entry and operational overheads. We start by covering the key principles and benefits of Serverless, with specific reference to compute, database and media services.
We then explain why the demand patterns that TV services place on IT resources often make them well-placed to benefit from Serverless, drawing on the concept of Elasticity of Capacity.
Following an illustration based on a hypothetical PPV scenario, we conclude with an assessment of current limitations, possible future evolution, and relevance to the industry.
No comments yet