Monitoring IP streams

Broadcasters and cable and IPTV providers have come to rely on IP transport of television content for one primary reason: It's cheap. IP-based infrastructures are not exactly ideal for carrying video between components in a facility, but they are much cheaper than other alternatives. They also provide suitable capacity, flexibility and interoperability, and they handle longer distances well enough to be an economical solution.

The combination of the IP infrastructure and MPEG standards makes flexible, cost-effective delivery of DTV content possible. The health of the IP network is key to the delivery of audio and video to the customer, and when tracking down impairments and their root causes, broadcasters can learn from analysis of the IP layer. After all, as content flows through a facility, it spends time as packets of data in an IP stream.

The Media Delivery Index (MDI) is one tool that gives broadcasters and service providers a measure for monitoring live video over IP flows. With knowledge of the timing and loss of each video stream throughout the network, the operator can target issues in the IP network and address failures. It's a mistake, however, to assume that if the IP network is OK, then video is OK as well.

Simple IP network monitoring may explain why data didn't arrive at its target destination, but analysis of the MPEG transport stream encapsulated into IP packets is critical to determining the “correctness” of DTV flow, which in turn is vital to the quality of the viewing experience. Analysis of the MPEG layer presents a significant challenge, but the ability to look deep into the video transport layer and examine all of its characteristics is essential to gaining valuable insight into service impairments.

Evolution of DTV monitoring

DTV systems are complex, comprising a variety of components that must connect and interact in a specific way to be successful. Monitoring the DTV transport stream and how well it meets industry specifications is also quite complex, but necessary for maintaining a high quality of service (QoS).

Because each receiver type responds differently to issues within the transport stream, the most reliable means of analysis is through comparison to set recommendations for various stream elements and parameters. Simple monitoring of a channel using a set-top box and TV will only demonstrate particularly bad problems with the transport stream and offers no hope of catching errors before they impact the viewer. The development and continued refinement of recommendations as to which parts of the stream must be verified has given broadcasters a far more sophisticated and effective technique for monitoring.

Initially, there were standards that dealt with how MPEG functions, but little practical help existed in applying those standards within a monitoring scheme. To address this type of problem for DVB streams, TR 101-290 was developed by DVB for the purpose of measuring transport stream parameters uniformly and according to three priority levels.

The ATSC built upon the concepts in TR 101-290 and created ATSC A/78, which takes a less black-and-white approach to error prioritization, establishing “gray scale” error conditions that may be filtered according to importance and impact on the viewer experience. ATSC A/78 served as the foundation for SCTE-142, which added support for cable specific parameters.

Both A/78 and SCTE-142 use five levels of error severity to support a practical layered approach to stream monitoring. Because digital video is complex and because of how the standards are written, it's rare to find a moment when there is no fault within the stream. Relatively insignificant errors crop up all the time; A/78 and SCTE-142 allow operators to set up intelligent filters and thus reduce the overall number of alarms. With this capability, it's easier to identify and localize critical errors while logging the rest for later observation or maintenance.

A/78 and SCTE-142 monitoring model

A/78 and SCTE-142 name seven categories of transport stream error type: PSI errors, out-of-band table errors, in-band table errors, PSIP errors, timing and buffer errors, consistency errors, and general errors. To these errors, the standard applies five levels of severity: transport stream off-air (TOA), program off-air (POA), component missing (CM), QoS and technically non-conformant (TNC). Table 1 on page 16 illustrates how a variety of program association table (PAT) error conditions, with differing levels of severity, are rated according to these recommended practices.

Whereas TOA, POA and CM typically reflect total video and audio loss and, as a result, demand immediate action, QoS errors are less serious in that the broadcast may be viewable but flawed. In this case, parameters are out of specification to the extent that a significant percentage of receivers likely will display compromised audio and video. A common example is when the PAT cycle exceeds the specification, leading to slow tuning from channel to channel. A TNC error violates the letter of the standard but has little practical impact on the viewer experience.

Internal monitoring schemes

The primary goal of monitoring is to catch errors before they cause noticeable degradation to or disruption of viewers' service. Thus, operators must make sure they are monitoring video streams at key points in their infrastructures. Ideally, every stream sent out to viewers should be actively monitored all of the time. As A/78 or SCTE-142 is increasingly built into monitoring equipment, and as that equipment evolves to handle a much larger volume of video flows at a reasonable cost, operators can benefit from implementing monitoring systems at multiple points.

On one end of the spectrum, broadcasters and service providers can use single-point verification for reference purposes, as in the simplified cable system illustrated in Figure 1 on page 18. Any type of plant benefits from more extensive monitoring, which allows the operator to identify errors as soon as they arise and helps engineering staff to localize problems quickly and resolve them before they compromise viewer services. By placing sufficient monitoring equipment throughout the plant to allow for a reasonable level of localization, the operator can realize good troubleshooting capabilities at a moderate cost.

At the other end of the spectrum, facilities seeking to establish maximum monitoring in a highly localized environment can place monitoring equipment “behind” every component that might impact the transport stream.

Monitoring external streams

While monitoring DTV transport streams within a broadcast or cable plant is indispensable to proactive maintenance of high-quality service, monitoring outgoing and incoming streams is valuable in DTV carriage auditing. Because most television viewers receive a broadcaster's service through a cable provider or other downstream infrastructure, auditing plays an important role in ensuring the sustained integrity of programming. Figure 2 demonstrates the concept of carriage auditing in action.

The quality of the broadcaster's service, including its video quality, as delivered through a cable plant or satellite operator can be significantly affected by decisions made by the downstream operator with respect to rate shaping, metadata accuracy, etc. Cost-effective stream-monitoring systems provide monitoring, measurement and even recording of DTV streams to ensure their integrity, reliability and compliance with standards.

The data delivered through the stream monitor facilitates rapid identification of issues and their origin, simplifying troubleshooting and serving as a basis for negotiating service-level agreements between broadcasters and cable, telco and IPTV providers. A standards-based monitoring model not only enhances quality of service, but also supports a stronger business model for all partners.

Richard Chernock is the chief technology officer at Triveni Digital.