Cloud infrastructure has done for the video ecosystem what aviation did for travel, yet for some reason the cloud DVR as a next-generation technology seems distanced from its sisters in the cloud. A webinar from IHS Markit this week therefore caught our attention and Faultline buckled up for what was sure to be anything but a bumpy ride.
IHS Markit bemoaned a lack of cloud DVR instances at pay TV operators, blaming various nebulous factors like scalability issues and copyright law as reasons to clinging onto legacy DVR hardware.
The way we see it, consumer demand today is quite clearly driven towards easy-access on-demand content. Demand for storing recordings in a private cloud, only then to access content which in most cases requires arduously skipping through lengthy ad breaks, is not the next frontier for on-demand video. Yes, there are major storage and hardware cost savings to be made and therefore plenty of vendor opportunities in the cloud DVR market, but the end trend is the same – the DVR is on the demise. Cloud or no cloud.
There were, however, a couple of interesting customer case studies from Anevia during this week’s webinar, a company we have come to associate with ABR packaging, not cloud DVR deployments. CTO Damien Lucas reeled off Net+ in Switzerland, a pay TV operator with 80,000 cloud DVR users – from its 180,000 total subscriber base – churning out 20,000 cloud DVR recordings every day. Net+ has slashed its set top return rate 10-fold by switching to cloud DVR using a combined Anevia-HPE system, the Apollo 4200 Gen 10 high capacity server which was introduced about a year ago. The Apollo is a very dense appliance capable reaching up to 350TB to 400TB.
Anevia also boasts Grupo Telecom in Argentina with around 1 million active OTT video users from its 4 million total TV subscriber base. The pay TV operator has reportedly doubled capacity with half the rack space through the emergence of hyper-converged infrastructure – the central technology on show.
Hyper-converged infrastructure is heavily disrupting the industry, or so HPE’s Manager of Datacenter & Hybrid Cloud, Olivier Tant, attests. As traditional storage appliances becomes increasingly expensive, working with commoditized hardware today is a much easier option. As such, a hyper-converged approach claims to optimize all elements of storage and compute components to work together on a single commodity appliance from a single vendor.
On the software side, five specific cloud DVR best practices were highlighted, covering just in the packaging, optimized recording, ad insertion, disaster recovery and distributed storage. The first software point, just in time packaging, offers a clever trick for cloud DVR use cases – optimizing the buffer by removing the end portion of the stream and therefore freeing up storage space. The others are more self-explanatory, with ad insertion offering additional revenue streams and optimized recording meaning chunks are stored only once.
IHS Markit made a big splash of saying pay TV ARPU is on the rise and has been recovering nicely since the 2016 trough. We probed whether this covered the 300 global pay TV operators it mentioned earlier in the webinar or if this was at a select few tier 1s, but that bullet was dodged.
The term cloud DVR first featured in the Faultline archive back in late 2012, emerging just before CES 2013 when companies like Boxee and TiVo set out to change the world, followed closely by Pace. Anevia is an example of the evolution, combining the encoding and origin CDN components which are generally not technologies known to plug together comfortably, according to Lucas. “There are not many vendors offering both these technologies,” he claimed.
The webinar was clearly targeted towards a certain type of operator and our press presence unwanted, with our four or five questions left unanswered during the Q&A session and replaced with what came across as a pre-prepared set of questions.
Lucas brought up an interesting point though, questioning where the sky is in terms of maximum retention rate. Should vendors cater for operators wanting their subscribers to store content for 2 months, 6 months, or substantially longer? Shockingly, some operators are offering cloud DVR storage for as long as 6 years, which of course is great news for vendors charging for the privilege.
Yet with the pay TV emphasis on live channels along within a multi-OTT app ecosystem, there is a resultant rise in CPU capacity, and so surely it makes more sense for operators to reduce storage capacity and instead invest in something like Android TV operator tier? We cannot see the benefit of allowing subscribers to store hundreds of hours of content for years on end, whether on-premise or using hyper-converged cloud infrastructure.