|
Service providers are making the transition to IP infrastructures to deliver video. This move offers significant economic and operational advantages, but it also comes with challenges. One of the main challenges is to ensure that the quality of the video is sufficiently high to attract and retain subscribers. Put another way, the service provider must be able to deliver a superior quality of experience (QoE) to drive adoption of IP video.
When designing a network infrastructure to deliver IP video, service providers must consider three main factors: line-rate forwarding, quality of service (QoS) and high availability. These capabilities in turn dictate the bandwidth availability, packet loss and jitter of each video stream, the primary factors that determine video quality. The choice of network components must be driven by these considerations. In particular, high-performance routing nodes are the key ingredients in building an IP video platform.
Introduction
From most of its 70-year history, video has been delivered using analog technology. Analog video is simple and reliable, and drove the widespread adoption of television as a mass medium. But as the medium grew, its inherent limitations—bandwidth inefficiency, transport quality and video processing equipment expense—drove the migration to digital video.
Digital video was first introduced commercially in 1986. The predominant delivery technology today, digital video has a number of benefits. The quality of digital video is generally higher than analog, and as importantly, is consistent over a wide range of delivery conditions. Coupled with MPEG compression technology, digital video is highly efficient. And because the audio and video are in digital form, they can be transmitted in a variety of formats.
The Next Evolution: Packet Transport
Seeking more efficiencies and economies of scale, service providers are rapidly moving to packet-based delivery for video. Figure 1 shows a typical IP video network architecture.
There are a number of factors driving service providers to IP video. For starters, there’s money to be made in the short term. Service providers can potentially increase operating margins by converging multiple services onto a single IP network. Doing so avoids equipment duplication, reduces training and maintenance expenses, and simplifies overall network administration.
Looking down the road, service providers see opportunities to differentiate themselves by offering innovative value-added services (for example, Internet surfing on a TV set or combined voice and video applications), many of which depend on a unified network infrastructure. As providers track the evolution of a relatively untamed and unpredictable marketplace, their best bet is to be agile and innovative. Delivering all their services over a single converged IP network is part of that strategy.
The promise of transporting video over an IP network is high, but as with any technology innovation, there are challenges. To fully understand those challenges requires understanding the differences between IP video and traditional broadcasting.
In fact, there is no concept of broadcasting per se in an IP network, in the sense that a video stream is sent to every user on the network. Instead, IP video uses two fundamentally different transport models: multicasting and unicasting. Virtually all commercial deployments of IP video use a mix of these two models.
Multicasting
Multicasting is similar to broadcasting in that multiple viewers can look at the same video stream at the same time. However, unlike broadcasting, multicasting is selective: A multicast stream is only sent to those viewers who have “signed up” for it. Multicasting makes efficient use of the transport network bandwidth, since a single stream traversing the core network can be viewed by a large number of end users.
However, that efficiency comes at the cost of network complexity. Individual nodes in the network must be smart enough to decide which of the downstream nodes need to receive a given multicast stream. (Forwarding the stream to all downstream nodes would essentially create a broadcasting model and negate the inherent efficiency of IP transport).
Multicasting is the workhorse of IP video transport. Most viewers, most of the time, are viewing content that others are also viewing —scheduled sitcoms and drama shows, live sporting events, and news programming. For these kinds of content, multicasting offers a highly efficient and practical way to transport video.
Managing Multicasting in the Video Network
Multicasting requires network nodes to determine routing video streams so that they reach the end users who have ordered them. But how is this information communicated?
|
