Slashdot Mirror
Students Use 802.11g To Save Cable Industry
LiquidFun writes "Business undergraduates at UC Berkeley's Haas School of Business have written an e-business case for one of their case competitions that describes how to use 802.11g wireless technology to distribute cable content, both interactive and broadcast, throughout the home. They mention features like video-on-demand, cable gaming, etc. and even provide enough of the technical specifications necessary to start believing that this could work. They even make available their PowerPoint presentation that they presented to judges from both Cisco & Deloitte Consulting. I'd say a pretty good job for third-year undergrads."
Lines like these might perk the interest of regulators.
Ryan Fenton
I agree, nice to see the thought and work put into this, but their are some obvious issues.
1st, cable and DLS are equal once within the house. This would not give cable any advantage over DSL.
2nd, this does not affect those who use Satellite because cable is unwilling to service them, not incapable.
3rd, security will not exist. Remember, the demand for hacks will skyrocket once you put it into each house. plus its wireless. Thats just asking to be hacked... Cable companies will have to spend MUCHO $$$ to prevent this on a yearly basis. You know, like they did with the cable box upgrades every 2 years, only much more agressively.
4th, 802.x works nice in a ranch house, but it will not like going through floors. It will loose lots of bandwidth there.
If you ignored competition, I think this would be an excellent idea, but I do not see it as a competitive one.
Under the 802.11g wireless standard, which is capable of transmitting data at a rate up to 54mpbs
No, it isn't. Without any interference from other 2.4GHz devices, you can't really expect more than 22Mbps. And that is shared bandwidth - once you have two way connections between the AP and several client devices, they all share the bandwidth. If you have hidden node issues (one client can't hear one of the other clients talking to the AP) you have to go to RTS/CTS mode to avoid collisions which also decrease the effective throughput. There is also no proper QOS in 802.11/b/g, first come first served - on a wireless segment with moderate to heavy load, you can experience lag spikes if several stations wish to transmit at the same time.
The router will also broadcast a connectionless stream under the UDP/IP protocol to devices such as televisions. This connectionless stream will be broadcasted in real-time, continuously, whether a device is "listening" or not, so several devices (e.g. televisions) can simply listen in when they are turned on without having to send a special signal to the box.
Hold on. Will the cable box translate MPEG2 streams received from the cable side to video frames and broadcast those to 802.11g enabled TVs, or will it just broadcast the MPEG2 stream?
In scenario one - have they done the math on how much bandwidth this will require? Not to mention the cost of upgrading your TV to receive raw video frames over 802.11g?
In scenario 2 - have they considered the cost of adding the MPEG2 and 802.11g hardware to the TV? What about MPEG2 artifacts if some frames are lost?
When broadcasting over wireless, you have to take into account the maximum speed of the client with worst signal quality. You can't expect to use 54Mbps modulation if you want the TV at the other end of the house to receive the signal with low packet loss. You can't rely on always being able to use 54Mbps (22Mbps throughput) if you want reliable broadcast to other devices in the household.
What makes cable gaming different from existing gaming networks is that with existing gaming networks, all processing is done locally on the user's own machine. With cable gaming, the required processing is done by the company's machines. This minimizes the actual amount of bandwidth required to travel along physical cable lines, as well as negates the need for a game processor on the consumer end (i.e. a game console; Xbox, Playstation, etc.). The only additional hardware required on the consumer's end is a minimal amount of Random Access Memory (RAM) onboard the digital cable box that acts as buffer memory to ensure a smooth, seamless gaming experience.
Sending complete video frames through the cable net is somehow less bandwidth consuming than sending UDP packets containing the state of the game? Even assuming an MPEG2 video stream, I don't buy that without seeing hard numbers. I would also worry a bit about latency.
The idea to use available bandwidth on the cable to provide new services is intriguing, and should definately be explored. But I think that the engineering needed to make something like this work is a bit higher than what the paper assumes.
If J.K.R wrote Windows: Puteulanus fenestra mortalis!