In my article The Future of Telephony and Communications vs. Toll-Quality I outlined that many years of beating by cell phones have lowered our standards in what we should expect from our carriers for voice quality. In an interesting paradigm, as we’ve given up quality for convenience in voice, we’ve demanded more and more in our desire to have bigger, better, higher quality video. From adding High-Definition to traditional broadcast TV to digital cable to satellite this has been easily accomplished but there’s another beast on the horizon that isn’t going to be easily satisfied: Internet Protocol Television or just IPTV.

Even though video in a transport scenario calls for a slightly lower QoS demand than voice, it demands much more bandwidth. IPTV has to deal with this as this big internet thing we’ve all come to rely upon is carrying it all: web surfing, mail, voice, and video.

What does this mean? Are the networks, superstations, and premium services going to forego the big transmitters, cable systems, and satellite to start broadcasting everything over the internet? No – there’s just not enough bandwidth there to do it today even using multicast. It will all rely on two things: when you want to watch your shows and what quality you’re willing to endure while watching.

Let’s just examine a worst-case scenario of watching a program on your 50” progressive-scan 1920×1080 resolution HDTV. Between it being HD and that high resolution you’ve just cranked up a 24Mb/s video stream with the technologies available today. OUCH! There’s not a lot of home broadband that will support that kind of streaming! So what happens if there are 50,000 people trying to watch this same show at the same time? No need in doing the math because there’s just not that much bandwidth out there. This is where IP multicast comes to play. Imagine for the sake of reference that you were to watch some 30-minute video on your PC that needed a 100kb/s stream. If it was a popular network TV program such as a sitcom there would be millions of people sitting there queued up to watch it at the same time. Regardless of the relatively small bandwidth needed, it being multiplied by millions would never work if each watcher had a dedicated session all the way to the broadcast source. By using IP multicast the source could serve streams to the limit the amount of bandwidth needed to a remote point such as a college campus or MDU (or city, or state, or region THINK BIG!!!) where the stream bandwidth would be multiplied by the number of watchers in that network. In the conceptual diagram [below] I’ve provided, you’ll see a single streaming source server with 4 connections outward to primary queue servers. Each primary queue server is linked to 3 more secondary queue servers. Each secondary is distributing to 5 end devices each, all of which using this theoretical 100kb/s stream to watch the program. By using IP multicasting the source server is supplying 4 100kb/s streams totaling 400kb/s to be distributed while there are, in actuality enough program viewers to consume 6Mb/s in bandwidth. While this is a concept only and any useful IP multicast video architecture would be much greater in both size and complexity, this should serve as a guideline as to how the technology would need to work. There are many more facets to IP multicasting such as Internet Group Messaging Protocol (IGMP) however these are beyond the focus of this article and will not be discussed in depth here. The last thing I want to leave with you before we close the book on multicast-101 is the timing of the streams. As the source server from the above concept streams to its primaries, the primaries will queue up the stream for a short time before making it available downstream to the secondary queue servers. This will be repeated each time a server pushes a stream to a secondary, tertiary, etc,… etc,… therefore the further down the line you pick up the stream the later you can start viewing. This can be in terms of seconds or even minutes after the broadcast begins.

With a basic understanding of how the stream would get from the serving “Internet TV Station” to you we now have to look at what would have to happen to see this sort of IPTV gaining popularity. Naturally, most of us want to watch our programs the instant they are broadcast but what many people are doing lately (as is the story at my house) is recording the show with a TiVo or some sort of DVR/PVR device and watching it later. Those of us that use the DVR/PVR technology are already in the first stage of moving toward IPTV for this reason without even knowing it. So what if an Internet TV station (for lack of a better name) was to start offering your favorite drama or sitcom as an “on demand” event which became available on a certain date and time? This means it would be available to watch for the first time next Tuesday night at 7pm central so you could watch it then OR, connect afterwards, call up the show and watch it over and over again being able to pause, jump forward/rewind as you watch. Pretty neat huh? I think what we’ll see in the near future is more internet sites coming up with programming options for you to watch until there becomes a complete alternate television programming source and traditional TV stations will be competing for your attention.
Now the downside – picture quality. We’ve all seen videos on youtube.com and putfile.com that range in quality from decent to ok to OH DEAR GOD CAN THIS BE ANY WORSE WITHOUT USING CRAYONS. Stream size is directly proportional to video size (the size of the window on the video player you’re using) and video quality. To date I haven’t seen anything that didn’t have some pretty strong dithering and artifacts as I increased the window size to full screen while watching a streamed video - nothing to match the quality of the worst cable or satellite channel. That would be on a laptop screen or a desktop monitor no larger than 17” for which I’ve tested. What happens when we try that on a 35” TV or a 50” big screen? I think it would almost unbearable to watch a low bandwidth stream.

Necessity is the mother of all invention and just as everything else, IPTV will improve at time goes on. If it’s truly the next thing this development will come faster. A dedicated streaming codec matched with downstream queuing servers and a video client in the form of software for a PC or “set-top” appliance for the TV will appear – at any rate, IPTV could very well change the way we watch TV in the future.

What about today in the MDU environment? The model still fits to where you could interface a Cable TV head-end and provide basic cable services in a controlled environment; all that’s needed is a way to get the stream on today’s conventional TV set. Most newer televisions (HD mostly) have many different types on inputs including the ability to act as a monitor for a PC!

This entry was posted on Thursday, March 29th, 2007 at 4:21 pm and is filed under Technology. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.