Getting the Quality of Service you need for OTT success

Andy Burnett, ITV and Dave Clark, Irdeto

A Q&A with Andy Burnett, Director of Online Technology and Operations, ITV and Dave Clark, Director of Service Delivery, Irdeto

Why do OTT and Quality of Service [QoS] so rarely square up?
[Andy Burnett] OTT services are now coming of age – as they increasingly move into the mainstream, viewers’ expectations rise.  Today, there is little tolerance for slow delivery or viewing glitches and most viewers now expect to enjoy a “broadcast” quality experience not only on the big screen at home, but also on their personal devices.
This is one of our motivations for launching ITV Hub – a new streamlined platform that will better answer our viewer needs for easy access to our content from whatever device they like.
[Dave Clark] OTT services are just as their name indicates – they go “over the top” of traditional managed networks and are dependent on a patchwork of independent IP networks to reach the end-user.  Not only is there the issue of shared bandwidth and throughput, but also typically a greater number of separate systems that need to interact – especially compared to broadcast.

What is the impact of poor service quality on your business?
[Andy Burnett] To put it simply, any shoddy service will lose customers.  We are in a high-engagement business where viewers lose interest exponentially with every second of buffering– this requires a good level of control over the end-to-end delivery.  We’re not only talking here about broadband throughput and speed, but also understanding the impact of backend systems such as content preparation workflows for live and near-live distribution.
In the early days of OTT services, everyone was just focused on getting to market, getting their content out there, but today there is so much choice, so many options, having appealing content is not the only factor influencing our customers.  You may be able to draw in new viewers with a live sporting event or X-Factor, but they will not come back to your service if they find it under par.

Why is it so hard to achieve good QoS on OTT platforms?
[Dave Clark] In order for video content to get to the end-user, it has to jump through a number of hoops.  Typically for OTT delivery, the bottlenecks arise from one or more of the various integration points between the multiple systems that process the content; encoding/transcoding, packaging, encryption, storage [CDNs], playout, etc…
With any number of different data flows that the content can take depending on the overall architecture, each OTT service will need to be carefully monitored to identify the capacity bottlenecks and potential failure points.

If there can be no guarantee of QoS in an IP world – what are our options?
[Dave Clark] In order to counter some of the things that can go wrong, Irdeto has put in place a systematic supervision of all our OTT deployments through our Service Operations Center.  This SOC provides 24/7 network supervision for all our customers worldwide and implements an “early alert” mechanism that kicks in when any anomalies are identified, so that preventative action can be taken BEFORE the end-user is affected.
Over time, the data collected allows us to understand where the weak points in any deployment or workflow are and to adjust accordingly.  Our service dashboards allow us establish whether things are running as normal or if an incident needs to be escalated.
[Andy Burnett] Indeed, the combination of service monitoring and pro-active alerts is so effective, we’ve enjoyed 100% platform availability for 15 of the past 16 months (with 1 month at 99.95%).

Our A-B-C for ensuring a good end-user experience?
[Andy Burnett] We want our different OTT services to be “as good as telly”, so here are our top 3 recommendations for delivering a top quality service:
A. Monitoring for deeper understanding: As no two network delivery systems behave the same, in order to understand the ITV viewer experience, we need to observe that behavior.  This is by continually monitoring and testing our workflows – why there was latency between data centers A and B, where the recurring bottlenecks and potential weak spots are.  Building on this service delivery data, we can quickly pin-point any issues by keeping an eye on the higher-risk areas.
B. Pre-defined process for crisis management: We systematically go through a number of “controlled disasters” – we introduce failure to better understand our systems, processes and people. We determine with our partners the escalation points, responsibilities and course of action required on both sides for every potential crisis. Trialing the specific processes for different situations when we “break” the services intentionally, allows us to avoid panic and wasting time when incidents do occur.
C. Take ownership: We depend on our partners to communicate pro-actively on potential service issues, which will require having the right level of expertise and sufficient coverage available 24/7.  Irdeto’s SoC is a good example of a reliable service monitoring mechanism to ensure service availability.  With full time staffing and early-alert mechanisms, it allows us to act quickly and remedy issues before they affect our end-users.
We’ve seen that these A-B-Cs have really helped us to achieve as stable a service as possible, so that our customers get a good viewing experience and keep coming back for more.

For more information visit www.irdeto.com