In the rush for content, don’t overlook the importance of storage

The essential technology elements for cloud DVR or time-shifted TV deployment similar.

One of the most critical aspects is data storage capacity, along with the necessary levels of performance to cope with growing recording archives, VOD libraries, numbers of channels, and time-shift buffers.

Estimating the storage requirements is difficult given the number of variables, including the unpredictable nature of some cloud DVR content.

An additional impact to storage requirements is the number of layers, file formats, and resolutions saved per program. The storage for a single program can multiply exponentially as you add devices and profiles to the storage mix.

This is further compounded by the shift from multicast linear TV to unicast VOD and cloud DVR content, which creates significant performance challenges.

Unicast requires a separate stream for every user, even if they are streaming the same content. This is unlike multicast delivery, which reaches hundreds of homes with a single video stream. In order to maintain an acceptable level of quality of service, there must be sustained high performance.

Cloud DVR offers highly efficient processing that helps compensate for the heavy compute requirements that come with the potential 30-40% recording concurrency at peak viewing times. Playout concurrency is also much higher than typical VOD applications with cloud DVR typically spreading over 72 hours after the recordings. Massive throughput is required for cloud DVR or time-shifted TV services due to the number of potential parallel recordings.

As many more operators consider the transition towards the cloud, it’s quite clear that not one configuration - dedicated or virtualised infrastructure - fits all.

There are several factors involved, ranging from running a cloud DVR or time-shifted solution within a dedicated hardware infrastructure or leveraging a virtualised cloud environment.

There is also the challenge of selecting service offerings, applying regulations and coordinating these processes with their current solutions.

For smaller pilot projects or for tier 2 operations serving smaller footprints, reusing dedicated infrastructure can make sense to maximise performance and storage of existing hardware, avoiding the ‘cloud tax’.

However, for larger subscriber communities or for new services that are scaling up quickly, a virtualised environment brings operational efficiencies and faster time to market with automatic elasticity and scalability.

A sustainable future

A move to a virtualised approach also helps with future flexibility.

So, is it better to have shared, private or both? The answer to this question touches upon one of the most interesting challenges, which is legal rather than technical.

Depending on jurisdiction and content type, operators may be forced to make separate copies of each item recorded by a subscriber in the cloud DVR instead of just a metadata pointer to a single shared copy.

However, the requirement of private versus shared copy in major markets like the US and Europe is likely to undergo several legal changes over the next few years.

A number of precedents have already been set in the U.S. thanks to two groundbreaking copyright law decisions: the ‘Betamax Case’ (1984) and the ‘Cablevision Case’ (2008).

The former ruling stated that the making of individual copies of complete shows for purposes of time shifting does not constitute copyright infringement, but rather fair use. The court also determined that the manufacturers of home video recording devices could not be liable for infringement.

The latter case found that Cablevision’s Remote-Storage Digital Video Recorder (RS-DVR) did not infringe copyright owners’ public performance rights instead, it was deemed as a private viewing, as the only person playing back a recording was the person who made it.

Storage solutions 

Each and every copy saved must be stored, which creates a very complex storage issue, requiring a solution that can handle the scale and performance requirements of such a service.

The key here again is the flexibility of the solution owned by the TV operators.

Cloud DVR design best practice suggests that hybrid deployments with multiple schemes deployed simultaneously is the safest option. The possibility to dynamically change according to changes in the regulatory environment going from private only copy to allowing shared copy or even content owners allowing certain new contents to be only private copy seems to be the best option.

Furthermore, hybrid deployments allow the possibility to use techniques like the de-duplication to reduce storage needs, where multiple copies are deleted and only a master copy is kept; on playout, files are reconstituted from the master and then served.

The usage of time-shifted TV or cloud DVR is growing and subscribers are now demanding more from their TV operators as new OTT providers attempt to attract viewers away with aggressive pricing models. Pay-TV operators need to be ready to offer the best quality of service and features to retain viewers.

Cloud DVR or time-shifted TV service brings personalised content to them, when they want it.

But cloud DVR or time-shifted TV isn’t a trivial deployment and the challenges are considerable. Legal issues, storage concerns and performance requirements must all be considered.

While there are a number of advantages to leveraging conventional hardware-based deployments, the convergence of the internet and media worlds will enable opportunities within virtualisation and data driven experiences that will offer unprecedented levels of scale and depth of capabilities.

Pay-TV operators must determine the strategy that best applies to their current and future potentials. For larger enterprises, for instance, the combination of a public and private cloud as part of a hybrid architecture may represent the best formula for success within a fully cloud based structure.

Ultimately, choosing the correct platform that supports hybrid cloud scenarios will prove to be pivotal in modernising and automating software.