Switched video has gained a lot of interest over the last year as MSOs try to extend their available bandwidth to offer more content and services. MSOs in the U.S. are in varying stages of switched video deployment. While most MSOs are facing bandwidth constraints, different MSOs have different strategies to enhance their bandwidth, based on their current state of network and alternatives considered for bandwidth reclamation. While MSO2[1] (Table 1-1) has been very aggressive with switched video market trials, some other large MSOs are primarily focusing on digital simulcasting and analog reclaim for reclaiming bandwidth. Whereas switched video is but one option for bandwidth reclamation, another option is to migrate analog programs to a digital line-up and to reclaim analog bandwidth in that manner.

1.2 Switched Video Architecture

A primary attraction of switched video is that it does not require truck-rolls[3] or set-top box replacements (for digital subscribers), and provides an incremental upgrade path to digital simulcasting that is already deployed in most networks. As shown in the list below, the switched video upgrade requires five incremental components.

a. SDV server - Collects channel change requests from STB client and commands the G-QAMs to switch IP multicast streams.
b. Master Session Resource Manager (SRM) - Allocates bandwidth in QAM between VOD and switched services, and enables existing VOD QAMs to be pooled and shared between VOD and switched video services.
c. STB client - Captures user channel change clicks and requests SDV servers. It also receives downstream commands from the SDV server to tune to the appropriate channel that matches the requested program.
d. G-QAMs - Narrow-cast G-QAMs need to be deployed at the nodes, to switch between multicast IP streams, as driven by Master SRM and SDV server. The G-QAMs have a GbE transport interface to join IP multicast streams.
e. Clamper - Typically, broadcast streams are variable-rate encoded and need to be clamped to a constant bit-rate stream (e.g. 3.75 MGps) before they can be switched.

Alternate VBR-based architectures for switched video are being proposed, to derive further bandwidth efficiencies by statistically multiplexing switched streams at the edge; however, products for such architectures are not expected until Q1/07 or later.

Conclusion — Executive Overview
The U.S. MSOs surveyed will use SDV mostly to save bandwidth for offering Long-Tail/Niche programming and HD. They also are considering on trial other bandwidth reclamation or preservation approaches including digital simulcasting statistical multiplexing wholesale conversion to digital from any addition of bandwidth and node-splitting (covered in the full version of this report); U.S. MSOs do not appear overly stressed about this issue, but clearly are taking aggressive steps to address a potential bandwidth shortage to compete with the likes of Verizon and its FTTP rollout strategy. Addressable advertising, while not a top priority for SDV, could come into play in a significant way by 2010.

Footnotes
[1] As a courtesy to survey participants for this study, each of the seven MSOs in this study is referred to by number.
[2] On average, digital subscribers make up 40-50% of all subscribers; yet, all subscribers have to be passed on a node to node basis due to the intermixing of analog and digital subs in each node.
[3] A non-responding Set-top Box (STB) (due to faulty splitterss inside the home or other reasons) may require a truck-roll; but this cost is ignored, as it is expected in <2-3% of the STBs in the field.