Cable’s Looming Challenges Delivering Video Over IP

Mon, 08/31/2009 - 8:00pm
Robert Winters, CMO at Shenick Network Systems

Can cable learn from how the telcos navigated this unchartered territory?

Delivering video over IP is akin to the pilgrims entering America for the first time. For the first group of pilgrims, life was tough. Subsequent settlers learned from their predecessors’ mistakes and successes and, while life was not a bed of roses, it was certainly a lot easier.

Although cable and telco companies are coming to the problem of delivering video over IP from different angles, both are facing the same issues. The telcos were the first to pioneer this territory, and – I believe – the cable operators can learn a lot from the trials and tribulations of the telco industry.

Telcos went from having voice nailed down to adding video and data, but video presents a different set of challenges, among them delay, packet loss and black screen cut-out. Many cable MSOs are now in the process of delivering video over IP as a solution, but it presents many challenges. So what should the cable operators be looking to learn from their competitors?

There is a catch: Although video represents a great opportunity and traffic is growing every month, at a fundamental level, IP networks were not designed for large quantities of time-sensitive video content. The telcos looked at a number of inherent issues in delivering video over IP infrastructure that presented them with some real challenges:

  • Performance of features and functionality – TV is an always-on service: You turn it on and expect it to work instantly and perfectly. Telcos needed to look to issues such as channel change rates in multicast-based broadcast IPTV. How quickly could subscribers change channels in congested scenarios, such as the ‘half time’ game surge in join/leave requests? Quality of experience (QoE) is obviously high on the list of end-user demands.
  • Scalability – Telcos looked at what happened when services started to scale to hundreds of thousands and moved into millions of users: Could their IP networks handle that load? Of course, adding users is the key to profitability, but poor QoE leads to churn, often resulting from inappropriate plans to scale.
  • Video media quality – What happens with the media? Is the video quality acceptable in small and large scale? What happens when users join and leave? Is video lost? Does the end user get good quality of service (QoS)?
  • Mixing the traffic – When you add the full multiplay mix of IPTV, VOD and VoIP and introduce additional data traffic such as P2P, what happens? Do all of these interact well together? How does the multiplay mix work for multicast IPTV, as well as unicast services like VOD? And more importantly, how does each individual application work for the end viewer when delivered with other services?

Telcos used a combination of test, measurement and monitoring to get the answers. The combination allowed operators to determine where issues were and gain insight into the performance and quality limitations of their IP-based services – how they were impacting business, and ultimately how to solve them. It provided them with a solution so that each and every application they offered – video, voice or data – behind every customer gateway could be offered to the level of service users expected. Telcos could fundamentally predict the QoS and QoE for every user and application.

Companies that specialized in IP test, measurement and monitoring worked with a lot of the telcos in the early days of IP video. In a test environment, products like Shenick’s own diversifyEye were being used to identify what each and every customer’s QoS and QoE was, right down to each application flow.

In a household where the kids are watching IP-based TV, a parent is working online and making VoIP calls, and another family member is watching a movie through a VOD service, test and monitoring systems could help the telcos predict each individual family member’s QoE and QoS – not just for each application type, but for each individual experience.

These same capabilities can be used by telcos to identify the top IP video quality issues leading to churn:

  • Black screen detection
  • Frozen frame detection
  • Audio cut-out detection
  • Poor IPTV channel change times

So what correlation can we draw for the cable environment? For a very long time, MSOs delivered video using a single pipe, but now that they are moving to the IP world, adding all of the services together is akin to the shock of an only child having three siblings born on the same day. In the IP world, video has to share the infrastructure with voice, data and other kinds of video traffic, such as broadcast IP video (using multicast delivery) and VOD (using unicast delivery). Added to this, there are all kinds of data and Web traffic such as P2P and YouTube.

Despite this, it makes sense for the cable operators to move to IP. Although challenging, the IP world offers a huge amount of flexibility and cost-effective implementation – that’s what the cable MSOs find attractive. With IP video, it’s easier to add extra channels; add more interactive, revenue-generating services; and, when done properly, more IP VOD, which is far more sophisticated than the typical time-shifted services offered today.

MSOs are already using limited IP in distribution of video, in the core of the cable network, such as multicast IP video delivery. But the bigger challenge for them is moving from an IP delivery mechanism right out to the subscriber. That’s where the rubber hits the road. It involves putting a new infrastructure in place, which has uncertainties about performance and quality, and these are the issues that need to be tested and addressed. The introduction of DOCSIS 3.0 at least allows for the introduction of much higher bandwidth per subscriber through channel bonding, providing a much cleaner environment within which to offer multiplay IP video services.

The capability exists to predict QoE and QoS in areas such as vendor selection or performance testing of the underlying cable infrastructure, such as at the cable modem. But the service and infrastructure still needs to be tested to extreme environments, particularly with mixed triple-play scenarios.

So what’s the advice for the cable industry?

Two things:

  • First, MSOs should be learning about a shared multiplay pipe. They need to be confident that they can offer a multiplay environment. They also need to be sure about QoE – it’s directly proportional to the level of churn experienced, but on the positive side, it adds to the profits they can make. The more user-sensitive applications that get put on the network, the more the quality side needs to be seriously looked into. In Europe, some telco operators have lost 15 percent or more of their customers because they couldn’t get IP video QoE right. Customer acquisition costs are often upwards of $250 a customer.
  • The cable industry should not underestimate the value of test and measurement. With the introduction of new high-bandwidth, real-time video services that stress IP network infrastructures, monitoring solutions become critical. High end-user expectations for video quality, especially for HD combined with the negative impact of delivering poor-quality services, can result in churn and costly new customer acquisition if not done properly. Cable companies should adopt test, measurement and monitoring as the very first step when rolling out new services or infrastructure – not as an afterthought – to ensure that when the service reaches the end user, it meets not just the end user’s expectations, but also every end user’s expectations.

For MSOs, delivering IP-based video services is fraught with risk, but the rewards are there for companies with the right approach to ensuring quality.


Share This Story

You may login with either your assigned username or your e-mail address.
The password field is case sensitive.