ANSI SCTE 168-7:2017 pdf free download – Recommended Practice for Transport Stream Verification in an IP Transport Network
4.0 INTRODUCTION
One of the challenges to detecting errors in the transport of MPEG streams is that these packets are carried in UDP, which is essentially connectionless. In order to determine that there is a failure there has to be a known metric by which to measure the traffic. In an IP Network there are specific network events that could lead to impairments or service degradation that would cause a perceived problem for the end user of a specific service such as watching a Video On Demand television program. Work has been documented in SCTE 142 [1] to detail media impairment severity levels specific to the connection between the emission remultiplexer and the QAM modulator. Recommended Practice SCTE 142 specifically refers to the MPEG Transport Stream and is independent of the underlying distribution system. Rather than using traditional ASI transport technologies, MPEG2 Transport Streams can be carried between the Head End System and a downstream Modulator over an IP-based network infrastructure such as that shown in Figure 1 with a simplified network view shown in Figure 2. When a packet network is used as the distribution system it could be non-deterministic and may cause errors that are reported at the transport stream layer when the SCTE 142 Recommended Practice has been implemented.
To ensure a deterministic model for the packet network, the network must adhere to a set of IP metrics that induce little or no impact to the MPEG transport stream being carried across the network. Because the type of media data in any given IP packet varies widely the effects of a single lost IP packet will vary. Most deployed networks do not prioritize specific packets based on content or handle packets with different content differently, and since any IP packet is expected to carry any mix of video, audio, or control information (PSI/SI tables), a single packet loss might affect video, audio, or the ability of a decoder to make sense of the data being received. With every packet loss carrying the likelihood for audio and/or video corruption, it is critical that an operator be able to detect all packet losses on all flows throughout the network and be able to determine when and where such losses occur for quality assessment and fault isolation maintenance. However, even though every IP packet loss is important to identify it is almost impossible to determine what type of media data was lost with that packet and thus very difficult to map a failure in the IP network to a specific MPEG error classification.
Unicast is a single destination model where the destination address denotes the end receiver and is matched to a single next-hop egress interface. Unicast addresses are used for one-to-one communication. Multicast is a one-to-many model where the destination address is not the end receiver, but instead denotes a multicast distribution tree with one or more egress interfaces. Unicast and multicast, however, both contain a forwarding plane and control plane even though they are unique in terms of routing topology creation and packet forwarding. Correct network setup and addressing schemes are basic network configuration concepts that are outside the scope of this document, but are required for the correct transport of packets.