Improving QoE for IP Video Services

The boom in OTT and TV Anywhere services is underlined by rapid growth in IP video transmission at all stages of the content lifecycle, and this is expanding greatly the scope and demand for Quality Assurance (QA) products. Even leading proponents of OTT services still admit there is some way to go to provide acceptable Quality of Experience (QoE) for high-definition premium content over unmanaged networks in particular.

“One of the main obstacles to OTT is the lack of a great user experience,” says Helge Høibraaten, CEO of Vimond Media Solutions, a spin-off of Norwegian commercial TV station TV 2, which is commercialising its OTT broadcast platform internationally.

Speaking at a conference during the recent IBC exhibition in Amsterdam, Høibraaten indicated that an OTT platform was defined by the quality it delivers and must meet the needs of all devices including tablets, PCs and smartphones. Vimond itself has only just extended its applications suite to Apple iOS devices (iPad and iPhone), Android and Windows phones, in addition to Windows desktop PCs which it already supported. The message for vendors of OTT platforms, and for the services that run on them, is that they should only embrace new device types when acceptable quality can be guaranteed.

The definition of acceptable quality is admittedly rather subjective. It is certain, though, that IP networks are creating new challenges for providers of QA video products. These vendors have been extending their portfolios to tackle video delivery over both managed and unmanaged IP networks, with various announcements made at IBC.

While unmanaged networks including the Internet pose the greatest challenge, even managed IP networks require careful handling to avoid packet loss and latency resulting from congestion within the infrastructure. This can happen because unlike traditional broadcast networks, IP infrastructures do not have fixed end-to-end paths and have no pre-determined transmission times for each IP packet. It is possible for more packets to enter the network than can be delivered within an acceptable time frame, leading to congestion and either dropped packets, delays, or both. Either of these can cause loss of quality on receiving devices.

The remedy is to apply traffic shaping, which involves holding up IP packets that are less critical or which can afford a little delay in order to preserve capacity for the most important packets. This can be performed at the point of entry to the network or within the network by routers themselves or other dedicated devices, and the key with managed networks is that operators can control the traffic shaping process better. Potentially, packet loss can be eliminated and latency kept within acceptable limits, according to Per Lindgren, VP Business Development and Co-Founder of Net Insight, the Swedish-owned vendor of the Nimbra IP media transport platform. Net Insight tackles the managed IP quality issue by breaking the network down into separate segments and applying QoE mechanisms including traffic shaping to each.

The first step is to ensure that the routers themselves do not create problems under congestion by dropping packets as they pass through, so Net Insight has applied traffic shaping at this level to ensure this does not happen. “By traffic shaping even inside our MSRs (Media Switch Routers), we can traffic shape down until we ensure we do not lose any packets there,” says Lindgren.

The next step is to address the links through the core network between the routers and ensure that the QoS needs of each individual service are met. “Traditionally telcos have not been treating media traffic as a special service,” says Lindgren. “So we propose building service aware media networks. MSRs aggregate traffic so that the core network (provided by a telco) only handles aggregated flows rather than individual services. Our MSRs then handle the different protection needs of each service, and can add QoS enhanced links inside a media service network rather than just at the edges.”

In this way, by addressing both the routers and links between them separately as part of a coordinated traffic management approach, the network can achieve much higher levels of quality. Even then, though, the possibility of packet loss or delay cannot be discounted, and so the third element of Net Insight’s QA strategy is to monitor every link. “We can do continuous real-time monitoring of traffic between MSRs and see any packet loss sent between one MSR and another,” Lindgren explains. “That makes it much easier to troubleshoot.”

Within unmanaged IP networks, on the other hand, it is impossible for broadcasters or operators to do either traffic shaping or performance monitoring since they do not own the infrastructure. This is an increasing issue with the growth of cloud-based services where the infrastructure is normally owned and managed by a third-party with video delivered over some Content Distribution Network (CDN). In that case there is an apparent black hole between the cloud and the end user, making it difficult for a content provider to know what quality the customer is getting.

Another Swedish vendor specialising in distributed video delivery, Edgeware, has tackled this problem with its Convoy VDN, which is software operating within the company’s Distributed Video Delivery Network (D-VDN) platform. Announced at IBC, this operates by combining the receiving device’s capability with the QoS known to be provided by the delivery infrastructure, according to Edgeware’s Chief Marketing Officer Duncan Potter.

The point is that CDNs usually operate via adaptive streaming protocols to improve network efficiency and performance, breaking video up into multiple small file chunks that can take different routes before being reassembled at the destination. The network detects each user’s CPU capacity and bandwidth continuously and adjusts the quality of the stream in real-time to ensure that QoE is always as good as it can be at that point in time. But breaking up video into chunks does make it hard to monitor what is going on within the CDN, and this is the problem Edgeware has addressed with Convoy VDN. “As we are a network device we can see what is going through,” said Potter. “We work out what is sent, collect statistics via a central reporting engine, and that is integrated with the higher level CDN management system.”

Such measures may help ensure optimum quality when a service is working normally but do not cater for major outages within the infrastructure. While IP networks are becoming more reliable, there is rising dependence in an increasingly global content market on external communication links that may be unreliable. This is a particular problem for the growing number of niche and ethnic services that have a global audience distributed across numerous, often small, communities around the world.

Such ethnic services can be lucrative, with high profit margins for operators because consumers are prepared to pay a premium or a separate subscription to receive them, but the total revenue in a given region is usually relatively small. This means operators cannot afford to spend too much capital on protecting against failure of the service in a region beyond their control, according to Danny Wilson, CEO of TV performance monitoring vendor Pixelmetrix. “Typically if an operator imports content from, say, India, they are vulnerable to loss of signal from Delhi,” he points out.

Pixelmetrix is tackling this with software announced at IBC that enables its DVStor recording and playback platform to perform disaster recovery and start playing out the content in the event of an outage. “We are recording what is going on at a downlink coming in from overseas and have integrated this with our test and measurement devices,” says Wilson. “Then if there is any interruption, the sensor detects that input signal is lost, and this DVStor solution can then provide back-up recovery on a real-time basis.”

This, in effect, is a cloud-based disaster recovery service and could be incorporated within IP-based delivery infrastructures. It highlights the growing scope of Quality Assurance, bringing together elements of disaster recovery, troubleshooting and performance monitoring within an overall QoE package.

By Philip Hunter, Videonet