The amount of video flowing across networks is growing rapidly. As broadcasters make the transition from a terrestrial to an IP-centric workflow, the ability to deliver video to IP centric devices, mobile and on-demand platforms is dependent on IP networks. Yet, the storage and processing of hundreds of thousands of hours of new material (and the millions more within video archives), poses a significant challenge to broadcasters.
More content, more data
Access to video via OTT and VOD platforms has reached an all-time high, in response to viewer appetite for extended catalogues of content. Last year, for example, BBC iPlayer opened up access to over 450 BBC archive programmes to UK viewers, from 1946 to the present day. Operators are now required to dedicate more technical resources and infrastructure to maintain these seemingly endlessly growing archives. A single piece of content might need to be transformed into 20 separate files to make it suitable for the various combinations of device, encryption and streaming technology. Add to this, the localisation for different languages, audience age groups or localised compliance regulations.
Recent evolutions in media-software are enabling a huge range of tools and services with the capacity to capture manage and analyse video of hundreds of petabytes per deployment. From applications, such as huge VOD libraries and large-scale cloud DVR, these solutions are opening a window of opportunity for the industry to deliver greater amounts of content to meet growing consumer demands.
When ‘big data’ becomes ‘big video’
The term ‘big data’ is often used to refer to unstructured, semi-structured and structured data that the commonly used software tools found difficult to process and generate insights. Analysing complex video is a major challenge today; the cost of capturing, curating, managing and processing video is often perceived to be too high to justify.
Although not a widely recognised term, we could describe this analysis as ‘big video’ analytics. According to Ericsson Mobility Report research, in 2017, video content accounted for 55% of all mobile data traffic, which will increase to around 75% of mobile data traffic by 2023 – making up more than 80 Exabytes per month. However, most of this information is discarded as the majority of the content is not stored anywhere and simply disappears seconds after being aired. The goal of ‘big video’ is quite simply to be able to capture, store and analyse as much of this content as possible to be able to keep relevant information and knowledge for future use.
If we look at the industry over the last 10 years, we can see that telcos, MSOs and the media are best equipped to help deal with this transition, due to their experience in media-video capture, delivery and management. For example, telcos are already addressing the complexity of video and have solved many of the issues relating to economical media delivery at scale.
Data requirements are increasing
At a time when video content is growing exponentially, many service providers are now following suit by turning to the cloud in an effort to tackle today’s data challenge. At a time when cloud DVR, on-demand and time/place shifted TV services are increasing rapidly, scalable and simplified architectures are crucial in helping these service providers to meet the needs of the TV Everywhere generation.
Beyond CAPEX and into economies of scale
However, the big video challenge is not just about scale. Video-based services are increasingly powered by lots of metadata. Everything from recommendation engines to service delivery and optimisation requires data to be gathered (often from multiple sources) and processed to gain some form of business insight. Again, cloud-based options provide a natural destination for the collection and processing of data, which has the benefit of utilising the same infrastructure as other video and service delivery elements.
The notion of big video should be considered as part of a wider and more long-term trend across the industry. For several years, the broadcast space has had to invest deep CAPEX into moving workloads into IP based systems to meet new on-demand distribution requirements. The demand and even the timescale from transmission to availability on secondary platforms has continued to shrink.
To improve agility and ultimately reduce costs, more elements of content production and distribution are heading into the cloud. From capturing rushes on set, to video FX and even entire SVOD services that live entirely on cloud – the desire to keep workloads in the same place and avoid ingress and egress charges for data across network makes the cloud more compelling. Each process can now be carried out in the same cloud using the same flexible and virtualized servers.
Embrace the big video opportunity
For operators keen to generate engaging services that retain and grow diverse customer communities, big video offers significant opportunity. As areas such as targeted advertising start to grow, VOD and streaming services are now looking at server-side ad-insertion of highly personalised messages. This is a process that requires both analysis of subscriber data and cloud-based capacity to quickly personalise these video streams and serve them to global audiences. Big video can also be used in other areas such as automatically generated content trailers and possibly in future AR/VR applications; Ericsson ConsumerLab’s ’10 Hot Consumer Trends for 2018 and Beyond’ predicted that ads using AR and VR will gain ‘app like functionality.’ Already more than half of current AR or VR users think ads will eventually replace the products being advertised. There are many possibilities… but what is clear is that the era of big video has arrived!