Why not multicast all of the way down?
May 8, 2007 12:00 PM
IPTV: So all of the channels that are aggregated are present down to the DSLAM?
MN: Correct. De facto from our technology, we multicast everything to every point in the network.
IPTV: Isn’t that a big departure from other IPTV architectures?
MN: Yes, it is. From our perspective, you are basically distributing all channels to the local consumption point, so you will avoid quality degradation at the start. Then the consumer is getting the response rate and the accessibility of the infrastructure because all of the content is pretty local, and the IP overhead is maintained at the access layer and does not have to travel through the core. That’s where you get the delays and the channel changing issues. That’s how we position our technology.
The way that we optimize and use the bandwidth, it’s efficient to do it and distribute it that way.
If it was just an IP infrastructure and you’re doing switching at Layer 3, then you get all of the overhead and bottlenecks associated with packetization. So from a video perspective, we’re saying, “Why not multicast all of the way down?”
IPTV: How is Net Insight addressing the VOD side of the IPTV business?
MN: From the implementations where we’ve had experience, we keep the VOD servers close to our nodes. So the latency between the VOD server is very minimal. In fact, they share the same infrastructure and sometimes sit in the same rack. So the communications channel between the two is a very short distance. You do not see the degradation associated with the latencies in the packet network because we are right next to the device itself.
As VOD plays out over one of our nodes, we take it to the access networks and to the DSLAMs directly. So it’s minimal impact, minimal infrastructure issues associated with that kind of configuration.
IPTV: Is there anything else you’d like to add?
MN: Virtual headends are a must for these guys. We’ve done ROI analysis, and the other offerings are very, very expensive compared to sharing infrastructure and sharing common VOD servers. So the rural telco market is very much going to be a community-based infrastructure. I cannot see projects like Lightspeed or FiOS taking place in that market. I think we are a good fit for these guys who want a lean, mean IP video experience.
IPTV: How far can you take the concept of a virtual head end? Do you see the possibility of a large number of rural telcos spread apart geographically forming into coalitions to share a headend?
MN: This week we spent time with three companies in the rural Minnesota area discussing how they can gain access to the infrastructure that’s already in place through fiber optic connection. This doesn’t just go into sharing content rights but sharing content administration systems, subscriber management systems, video servers and so on. You’re just adding subscribers.
IPTV: But is there a physical limitation, or is the speed of light the speed of light so it doesn’t make a practical difference how far the virtual headends are apart for the real headend?
MN: Correct. We’re looking at OC-192 links between these sites and off we go with 200 plus channels, and we just propagate across the U.S. geography by geography. Now obviously we’ll have to build in another headend somewhere so there’s redundancy in case a super headend fails another one kicks in, but then they are sharing infrastructure and it’s a per-subscriber fee rather than a CAPEX expense to build the infrastructure.
Tell Us What You Think!
IPTVU invites response from our readers. Please submit your comments to editor@broadcastengineering.com. We'll follow up with your comments in an upcoming issue.
| Want to use this article? Click here for options! |  |
